CustomerLifetime&SocialNetworkAnalysis.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/arpitamangal/customer-segmentation/blob/main/CustomerLifetime%26SocialNetworkAnalysis.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RxjZW8KHywfZ"
      },
      "outputs": [],
      "source": [
        "# Run below commands\n",
        "#!apt-get install openjdk-8-jdk-headless -qq > /dev/null\n",
        "#!wget -q https://archive.apache.org/dist/spark/spark-3.1.2/spark-3.1.2-bin-hadoop3.2.tgz\n",
        "#!tar xf spark-3.1.2-bin-hadoop3.2.tgz\n",
        "#!pip install -q findspark"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pyspark"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ALJlSDNjTEBY",
        "outputId": "4360ac71-249d-42dd-b0c3-89dab4e07bd0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting pyspark\n",
            "  Downloading pyspark-3.4.0.tar.gz (310.8 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m310.8/310.8 MB\u001b[0m \u001b[31m4.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: py4j==0.10.9.7 in /usr/local/lib/python3.10/dist-packages (from pyspark) (0.10.9.7)\n",
            "Building wheels for collected packages: pyspark\n",
            "  Building wheel for pyspark (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyspark: filename=pyspark-3.4.0-py2.py3-none-any.whl size=311317130 sha256=6ae312a3a96ca529d25d43eb2e6f8f0605b401c7a4f6d0e4bcc54a11f9665cba\n",
            "  Stored in directory: /root/.cache/pip/wheels/7b/1b/4b/3363a1d04368e7ff0d408e57ff57966fcdf00583774e761327\n",
            "Successfully built pyspark\n",
            "Installing collected packages: pyspark\n",
            "Successfully installed pyspark-3.4.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#!pip install pyspark\n",
        "#-------------------------------------------------\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark import SparkConf, SparkContext\n",
        "from datetime import datetime, date, timedelta\n",
        "from dateutil import relativedelta\n",
        "from pyspark.sql import SQLContext, Row\n",
        "from pyspark.sql.types import *\n",
        "from pyspark.sql.functions import *\n",
        "from pyspark.sql import DataFrame\n",
        "from pyspark.sql.functions import *\n",
        "from pyspark.sql.functions import to_timestamp, to_date\n",
        "from pyspark.sql import functions as F\n",
        "from pyspark.sql.functions import collect_list, collect_set, concat, first, array_distinct, col, size, expr\n",
        "from pyspark.sql import DataFrame\n",
        "import random\n",
        "import pandas as pd"
      ],
      "metadata": {
        "id": "JPJjoQYvTgRr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#import os\n",
        "#os.environ[\"JAVA_HOME\"] = \"/usr/lib/jvm/java-8-openjdk-amd64\"\n",
        "#os.environ[\"SPARK_HOME\"] = \"/content/spark-3.1.2-bin-hadoop3.2\""
      ],
      "metadata": {
        "id": "Gp4O-F8jIt6f"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#import findspark\n",
        "#findspark.init()\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark import SparkConf, SparkContext\n",
        "from datetime import datetime, date, timedelta\n",
        "from dateutil import relativedelta\n",
        "from pyspark.sql import SQLContext, Row\n",
        "from pyspark.sql.types import *\n",
        "from pyspark.sql.functions import *\n",
        "from pyspark.sql import DataFrame\n",
        "from pyspark.sql.functions import *\n",
        "from pyspark.sql.functions import to_timestamp, to_date\n",
        "from pyspark.sql import functions as F\n",
        "from pyspark.sql.functions import collect_list, collect_set, concat, first, array_distinct, col, size, expr\n",
        "from pyspark.sql import DataFrame\n",
        "import random\n",
        "import pandas as pd\n",
        "import re"
      ],
      "metadata": {
        "id": "1K7fLrV7IwWD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession\n",
        "from pyspark import SparkConf, SparkContext"
      ],
      "metadata": {
        "id": "c1Hu1adwSAlm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the spark session\n",
        "spark = SparkSession\\\n",
        "        .builder\\\n",
        "        .getOrCreate()"
      ],
      "metadata": {
        "id": "mbsS8I_pIwcw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xwPRbEXaIwjk",
        "outputId": "cd809159-030a-41f6-f49e-cd8c084aef52"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import udf\n",
        "from pyspark.sql.types import ArrayType, StringType"
      ],
      "metadata": {
        "id": "-v2IGnkbIwoh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Text Analytics"
      ],
      "metadata": {
        "id": "bUY6IJPIQqOb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "venmo_df = spark.read.parquet(\"VenmoSample.snappy.parquet\")\n",
        "\n",
        "# take a random sample of the Venmo DataFrame\n",
        "venmo_sample_df = venmo_df.sample(fraction=0.1, seed=42) # specify the fraction and seed as desired"
      ],
      "metadata": {
        "id": "2lEoEiteIwr7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "venmo_sample_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4wG3OgarIwua",
        "outputId": "4ec25d47-d1cb-48fa-b460-e62baf27af06"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "| 730075| 804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|\n",
            "|4682257|1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|\n",
            "| 241386|2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|\n",
            "|3393501|1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|\n",
            "|1864528| 437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|\n",
            "|5670003| 618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|\n",
            "|1168366| 735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|\n",
            "|1751344|2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|\n",
            "|3658258|3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|\n",
            "|4828390|4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|\n",
            "|8894083|2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|\n",
            "| 411633|4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|\n",
            "| 303413|3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|\n",
            "|1158024|3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...|\n",
            "|6638728|8528091|         payment|2016-02-23 10:26:27|      Parade day🍻🍻|      false|56cbc353cd03c9af2...|\n",
            "|3204155|4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|\n",
            "|7104084|7006896|         payment|2015-12-09 05:15:04|                Food|      false|56674858cd03c9af2...|\n",
            "|2674970|3691801|         payment|2015-06-19 06:05:46|               Foodz|      false|55834ecb7b9372746...|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# load the text and emoji dictionaries as PySpark DataFrames or RDDs\n",
        "text_dict_df = spark.read.csv(\"Venmo Word Classification Dictonary Word_Dict.csv\", header=True, inferSchema=True)\n"
      ],
      "metadata": {
        "id": "NLELAoYJIww4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "text_dict_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kZmd0-0eMqUO",
        "outputId": "2203f095-0049-4a57-d322-49c123610f6f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+----------+---------+------------+--------+-------+--------------+-----------+---------+---------------+\n",
            "|    People|     Food|       Event|Activity| Travel|Transportation|    Utility|     Cash|Illegal/Sarcasm|\n",
            "+----------+---------+------------+--------+-------+--------------+-----------+---------+---------------+\n",
            "|    friend|     food|    birthday|    ball|  beach|          lyft|       bill|     atm |      addiction|\n",
            "|friendship|      bbq|   christmas|    boat|  place|          uber|      cable|    bank |           drug|\n",
            "|      baby|     bean|       happy|     bar|     la|           cab|        fee|    cash |          wangs|\n",
            "|       boy|    latte|        bday|    book|  world|           bus|   electric|    money|           weed|\n",
            "|      girl|breakfast|     wedding|    club|  hotel|           car|electricity|     buck|           anal|\n",
            "|      help|   brunch|        xmas|    card|   trip|           gas|   internet|   wallet|             bj|\n",
            "|      like|   burger|     holiday|   dance|   vega|          taxi|       rent|   monies|        blowjob|\n",
            "|      love|  burrito|         hbd|football|  tahoe|          ride|       wifi|      tip|           boob|\n",
            "|       mom|     cake|   halloween|     fun|    nyc|       rental |    utility|   dollar|          booty|\n",
            "|      save|   cheese|thanksgiving|    game|     dc|         train|        tax|  payback|           blow|\n",
            "|     sweet|  chicken|    bachelor|    gift|     sf|          über|     refund|  lettuce|        cocaine|\n",
            "|      tank|   coffee|        gift|    golf| island|         ubers|      house|    watch|           cock|\n",
            "|     thank|     chip|    donation|    hair|   cabo|      gasolina|      billz|      hat|          dirty|\n",
            "|    thanks|  chinese|     charity|      pt|chicago|       airport|    deposit|  supreme|             dd|\n",
            "|       thx| chipotle|         aid|   movie| airbnb|          fuel|         pg|  bitcoin|          crack|\n",
            "|        ya|    cream|         nye|  lesson|travel |        cruise|         tv|robinhood|           bang|\n",
            "|        yo|   costco|   festivity|    park|boston |           air|      paper|  balance|           dank|\n",
            "|       lol|   dinner|      formal|   party| hostel|          limo|    comcast| expenses|           dick|\n",
            "|        jj|    drink|     shower |   poker|retreat|          bike|      light|   switch|           fart|\n",
            "|      lyfe|    drank|       photo|    pool|   nola|            bp|      phone|    extra|           fuck|\n",
            "+----------+---------+------------+--------+-------+--------------+-----------+---------+---------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "emoji_dict_df = spark.read.csv(\"Venmo_Emoji_Classification_Dictionary.csv\", header=True, inferSchema=True)"
      ],
      "metadata": {
        "id": "2CMj0l_8MrIu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "emoji_dict_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9bQC7e7YMrUf",
        "outputId": "918944aa-2a83-4394-f40c-218137334b20"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+------+----+--------+--------------+------+-------+\n",
            "|Event|Travel|Food|Activity|Transportation|People|Utility|\n",
            "+-----+------+----+--------+--------------+------+-------+\n",
            "| 🇦🇺|    🏔|  🍇|      👾|            🚄|    😀|      ⚡|\n",
            "| 🇫🇷|     ⛰|  🍈|      🕴|            🚅|    😃|     💡|\n",
            "|   🎂|    🌋|  🍉|      🎪|            🚆|    😄|     🔌|\n",
            "|   🛍|    🗻|  🍊|      🎭|            🚇|    😁|     📺|\n",
            "| 🇨🇦|    🏕|  🍋|      🎨|            🚈|    😆|     🔌|\n",
            "| 🇧🇷|    🏖|  🍌|      🎰|            🚉|    😅|      ⚡|\n",
            "|   🐉|    🏜|  🍍|      🚣|            🚊|    🤣|     💡|\n",
            "|   🎅|    🏝|  🍎|      🛀|            🚝|    😂|     💸|\n",
            "| 🇲🇽|    🏞|  🍏|      🎗|            🚞|    🙂|     💦|\n",
            "| 🇨🇳|    🏟|  🍐|      🎟|            🚋|   🙃-|     💧|\n",
            "|   🐰|    🏛|  🍑|      🎫|            🚌|    😉|     📱|\n",
            "|   🎥|    🏗|  🍒|      🎖|            🚍|    😊|     🏢|\n",
            "|   🍂|    🏘|  🍓|      🏆|            🚎|    😇|     🏣|\n",
            "|   👨|    🏙|  🍅|      🏅|            🚏|    🥰|     🔥|\n",
            "|   💪|    🏚|  🍆|       ⚽|            🚐|   😍-|   null|\n",
            "|   🎓|    🏢|  🌽|       ⚾|            🚑|   🤩-|   null|\n",
            "|   🎃|    🏣|  🌶|      🏀|            🚒|    😘|   null|\n",
            "|   🕎|    🏤|  🍄|      🏐|            🚓|    😗|   null|\n",
            "|   🕉|    🏥|  🌰|      🏈|            🚔|     ☺|   null|\n",
            "| 🇺🇸|    🏦|  🍞|      🏉|            🚕|    😚|   null|\n",
            "+-----+------+----+--------+--------------+------+-------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pyspark.sql.functions as F\n",
        "import re\n",
        "\n",
        "# Define a regular expression pattern to match emojis\n",
        "emoji_pattern = r\"[\\U0001F300-\\U0001F64F]|[\\U0001F680-\\U0001F6FF]|[\\u2600-\\u26FF\\u2700-\\u27BF]\"\n",
        "\n",
        "# Create a UDF (User Defined Function) to extract emojis from the description\n",
        "extract_emojis = F.udf(lambda text: ''.join(re.findall(emoji_pattern, text)))\n",
        "\n",
        "# Create a UDF to remove emojis from the description\n",
        "remove_emojis = F.udf(lambda text: re.sub(emoji_pattern, '', text))\n",
        "\n",
        "# Create a UDF to convert text to lowercase\n",
        "convert_to_lowercase = F.udf(lambda text: text.lower())\n",
        "\n",
        "# Add a new column 'emojis' containing the extracted emojis\n",
        "venmo_sample_df = venmo_sample_df.withColumn('emojis', extract_emojis(F.col('description')))\n",
        "\n",
        "# Add a new column 'text_only' containing the text without emojis\n",
        "venmo_sample_df = venmo_sample_df.withColumn('text_only', remove_emojis(F.col('description')))\n",
        "\n",
        "# Add a new column 'text_only_lower' containing the lowercase text\n",
        "venmo_sample_df = venmo_sample_df.withColumn('text_only_lower', convert_to_lowercase(F.col('text_only')))"
      ],
      "metadata": {
        "id": "CaXEDekjMrXN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "venmo_sample_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "43PUvstcMrbD",
        "outputId": "6b18f6a6-2e4c-446e-c388-6c0411bd3e26"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+\n",
            "| 730075| 804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|\n",
            "|4682257|1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|\n",
            "| 241386|2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|\n",
            "|3393501|1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|\n",
            "|1864528| 437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|\n",
            "|5670003| 618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|\n",
            "|1168366| 735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|\n",
            "|1751344|2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning|\n",
            "|3658258|3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|\n",
            "|4828390|4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|\n",
            "|8894083|2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|\n",
            "| 411633|4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|\n",
            "| 303413|3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|\n",
            "|1158024|3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|\n",
            "|6638728|8528091|         payment|2016-02-23 10:26:27|      Parade day🍻🍻|      false|56cbc353cd03c9af2...|     🍻🍻|          Parade day|          parade day|\n",
            "|3204155|4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|\n",
            "|7104084|7006896|         payment|2015-12-09 05:15:04|                Food|      false|56674858cd03c9af2...|         |                Food|                food|\n",
            "|2674970|3691801|         payment|2015-06-19 06:05:46|               Foodz|      false|55834ecb7b9372746...|         |               Foodz|               foodz|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.ml.feature import Tokenizer\n",
        "\n",
        "# Create a tokenizer instance\n",
        "tokenizer = Tokenizer(inputCol='text_only_lower', outputCol='tokens')\n",
        "\n",
        "# Apply the tokenizer to the dataframe\n",
        "venmo_sample_df = tokenizer.transform(venmo_sample_df)"
      ],
      "metadata": {
        "id": "zmzjZWRtMrd8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "venmo_sample_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LZMLA_cvMrgk",
        "outputId": "d7d56763-42a8-4108-8fff-a86f3554588e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|              tokens|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+\n",
            "| 730075| 804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|              [food]|\n",
            "|4682257|1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|   [, gypsy, things]|\n",
            "| 241386|2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|         [furniture]|\n",
            "|3393501|1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|         [bleachers]|\n",
            "|1864528| 437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|[the, decline, of...|\n",
            "|5670003| 618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|            [crepes]|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|\n",
            "|1168366| 735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|[shipping, and, h...|\n",
            "|1751344|2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning| [5/27, drycleaning]|\n",
            "|3658258|3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|            [closet]|\n",
            "|4828390|4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|        [u, trippin]|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|\n",
            "|8894083|2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|[outside, lands, ...|\n",
            "| 411633|4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|         [groceries]|\n",
            "| 303413|3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|             [chick]|\n",
            "|1158024|3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|\n",
            "|6638728|8528091|         payment|2016-02-23 10:26:27|      Parade day🍻🍻|      false|56cbc353cd03c9af2...|     🍻🍻|          Parade day|          parade day|       [parade, day]|\n",
            "|3204155|4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|[extra, cuz, i, l...|\n",
            "|7104084|7006896|         payment|2015-12-09 05:15:04|                Food|      false|56674858cd03c9af2...|         |                Food|                food|              [food]|\n",
            "|2674970|3691801|         payment|2015-06-19 06:05:46|               Foodz|      false|55834ecb7b9372746...|         |               Foodz|               foodz|             [foodz]|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Classify Venmo’s transactions"
      ],
      "metadata": {
        "id": "cgI9DkRgM97t"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Reshaping the text dataset:\n",
        "from pyspark.sql.functions import expr, col, lit\n",
        "\n",
        "# Rename the \"Illegal/Sarcasm\" column\n",
        "text_dict_df = text_dict_df.withColumnRenamed(\"Illegal/Sarcasm\", \"IllegalSarcasm\")\n",
        "\n",
        "# Reshape the DataFrame:\n",
        "reshaped_df = (text_dict_df\n",
        "    .selectExpr(\"stack(\" + str(len(text_dict_df.columns)) + \", \" + \", \".join([\"'\"+c+\"', \"+c for c in text_dict_df.columns]) + \") AS (category, value)\")\n",
        "    .orderBy(col('category'))\n",
        "    .filter(col('value').isNotNull()) # Exclude rows where \"value\" is null\n",
        ")\n",
        "\n",
        "# Show the reshaped DataFrame\n",
        "reshaped_df.show(truncate=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iq6RFLIBMrjY",
        "outputId": "e9d8d5ef-9c91-4446-c414-a936a87bce5d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+--------+\n",
            "|category|value   |\n",
            "+--------+--------+\n",
            "|Activity|gift    |\n",
            "|Activity|ticket  |\n",
            "|Activity|boat    |\n",
            "|Activity|book    |\n",
            "|Activity|club    |\n",
            "|Activity|card    |\n",
            "|Activity|dance   |\n",
            "|Activity|football|\n",
            "|Activity|fun     |\n",
            "|Activity|game    |\n",
            "|Activity|bar     |\n",
            "|Activity|ball    |\n",
            "|Activity|golf    |\n",
            "|Activity|hair    |\n",
            "|Activity|pt      |\n",
            "|Activity|movie   |\n",
            "|Activity|lesson  |\n",
            "|Activity|park    |\n",
            "|Activity|party   |\n",
            "|Activity|poker   |\n",
            "+--------+--------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Reshape the emoji dataset:\n",
        "\n",
        "from pyspark.sql.functions import expr, col, lit\n",
        "\n",
        "# Reshape the DataFrame\n",
        "reshaped_emoji_df = (emoji_dict_df\n",
        "    .selectExpr(\"stack(\" + str(len(emoji_dict_df.columns)) + \", \" + \", \".join([\"'\"+c+\"', \"+c for c in emoji_dict_df.columns]) + \") AS (category2, value2)\")\n",
        "    .orderBy(col('category2'))\n",
        "    .filter(col('value2').isNotNull()) # Exclude rows where \"value\" is null\n",
        ")\n",
        "\n",
        "# Show the reshaped DataFrame\n",
        "reshaped_emoji_df.show(truncate=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "04SmYbPHMrpt",
        "outputId": "d57dd805-815b-44b9-db15-3cbe0fbbfc7d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---------+------+\n",
            "|category2|value2|\n",
            "+---------+------+\n",
            "|Activity |🏅    |\n",
            "|Activity |🎿    |\n",
            "|Activity |⚽     |\n",
            "|Activity |🕴    |\n",
            "|Activity |⚾     |\n",
            "|Activity |🎭    |\n",
            "|Activity |🏀    |\n",
            "|Activity |🎰    |\n",
            "|Activity |🏐    |\n",
            "|Activity |🛀    |\n",
            "|Activity |🏈    |\n",
            "|Activity |🎟    |\n",
            "|Activity |🏉    |\n",
            "|Activity |🎖    |\n",
            "|Activity |🎾    |\n",
            "|Activity |🏏    |\n",
            "|Activity |🏑    |\n",
            "|Activity |🚣    |\n",
            "|Activity |🏒    |\n",
            "|Activity |🎨    |\n",
            "+---------+------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import array_contains\n",
        "\n",
        "# Perform the first join operation\n",
        "joined_df = venmo_sample_df.join(reshaped_df, array_contains('tokens', reshaped_df.value), 'left')\n",
        "\n",
        "# Perform the second join operation\n",
        "#joined_df = joined_df.join(reshaped_emoji_df, array_contains(reshaped_emoji_df['value'], joined_df['emojis']), 'left')\n",
        "joined_df = joined_df.join(reshaped_emoji_df, col(\"emojis\").contains(reshaped_emoji_df.value2), \"left_outer\")\n",
        "joined_df = joined_df.withColumn('category_class',\n",
        "                                 when(col('category').isNull(), col('category2'))\n",
        "                                 .otherwise(col('category')))\n",
        "\n",
        "joined_df = joined_df.withColumn('category_class',\n",
        "                                 when(col('category_class').isNull(), 'unclassified')\n",
        "                                 .otherwise(col('category_class')))\n",
        "# Display the resulting dataframe\n",
        "joined_df.show(50)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0roY_FodNGCK",
        "outputId": "94afff4f-4d3f-4b9f-e9d8-3c4fc5cbea6a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+\n",
            "|   user1|   user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|              tokens|      category|      value|category2|value2|category_class|\n",
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+\n",
            "|  730075|  804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|\n",
            "| 4682257| 1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|   [, gypsy, things]|          null|       null|     null|  null|  unclassified|\n",
            "|  241386| 2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|         [furniture]|       Utility|  furniture|     null|  null|       Utility|\n",
            "| 3393501| 1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|         [bleachers]|          null|       null|     null|  null|  unclassified|\n",
            "| 1864528|  437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|[the, decline, of...|          null|       null|     null|  null|  unclassified|\n",
            "| 5670003|  618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|            [crepes]|          null|       null|     null|  null|  unclassified|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|    Event|    🎂|         Event|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍳|          Food|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍰|          Food|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|     ☕|          Food|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍴|          Food|\n",
            "| 1168366|  735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|[shipping, and, h...|Transportation|   shipping|     null|  null|Transportation|\n",
            "| 1751344| 2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning| [5/27, drycleaning]|       Utility|drycleaning|     null|  null|       Utility|\n",
            "| 3658258| 3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|            [closet]|          null|       null|     null|  null|  unclassified|\n",
            "| 4828390| 4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|        [u, trippin]|          null|       null|     null|  null|  unclassified|\n",
            "|  179833|  942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|      foodz|     null|  null|          Food|\n",
            "|  179833|  942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|         za|     null|  null|          Food|\n",
            "| 8894083| 2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|[outside, lands, ...|          null|       null|     null|  null|  unclassified|\n",
            "|  411633| 4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|         [groceries]|          Food|  groceries|     null|  null|          Food|\n",
            "|  303413| 3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|             [chick]|          Food|      chick|     null|  null|          Food|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍦|          Food|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍩|          Food|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍪|          Food|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|   People|    👯|          Food|\n",
            "| 6638728| 8528091|         payment|2016-02-23 10:26:27|      Parade day🍻🍻|      false|56cbc353cd03c9af2...|     🍻🍻|          Parade day|          parade day|       [parade, day]|      Activity|        day|     Food|    🍻|      Activity|\n",
            "| 3204155| 4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|[extra, cuz, i, l...|        People|       love|     null|  null|        People|\n",
            "| 3204155| 4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|[extra, cuz, i, l...|          Cash|      extra|     null|  null|          Cash|\n",
            "| 7104084| 7006896|         payment|2015-12-09 05:15:04|                Food|      false|56674858cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|\n",
            "| 2674970| 3691801|         payment|2015-06-19 06:05:46|               Foodz|      false|55834ecb7b9372746...|         |               Foodz|               foodz|             [foodz]|          Food|      foodz|     null|  null|          Food|\n",
            "| 6523579| 1323619|         payment|2016-07-04 11:16:16|             Ran old|      false|5779e31023e064eac...|         |             Ran old|             ran old|          [ran, old]|          null|       null|     null|  null|  unclassified|\n",
            "| 3413708| 1589735|          charge|2015-02-05 08:02:14|          Trash bill|      false|54d2b30791bd05aa9...|         |          Trash bill|          trash bill|       [trash, bill]|       Utility|       bill|     null|  null|       Utility|\n",
            "| 3413708| 1589735|          charge|2015-02-05 08:02:14|          Trash bill|      false|54d2b30791bd05aa9...|         |          Trash bill|          trash bill|       [trash, bill]|       Utility|      trash|     null|  null|       Utility|\n",
            "|11339258| 1655285|          charge|2016-09-23 00:14:44|        October wifi|      false|57e4118423e064eac...|         |        October wifi|        october wifi|     [october, wifi]|       Utility|       wifi|     null|  null|       Utility|\n",
            "| 4226498| 4708480|         payment|2016-04-18 01:21:10|                  🍺|      false|5713d416cd03c9af2...|       🍺|                    |                    |                  []|          null|       null|     Food|    🍺|          Food|\n",
            "|  230898|  731601|         payment|2016-07-01 00:58:39|Reverse engineeri...|      false|57755dcf23e064eac...|         |Reverse engineeri...|reverse engineeri...|[reverse, enginee...|          null|       null|     null|  null|  unclassified|\n",
            "|  279563|  767893|         payment|2015-05-27 07:17:22|                Time|      false|55650d12d4aeb63c2...|         |                Time|                time|              [time]|          null|       null|     null|  null|  unclassified|\n",
            "| 3562883| 7420992|         payment|2016-09-29 00:16:21|     Daddy's wealthy|      false|57ebfae523e064eac...|         |     Daddy's wealthy|     daddy's wealthy|  [daddy's, wealthy]|          null|       null|     null|  null|  unclassified|\n",
            "| 6471126|10046265|          charge|2016-05-18 08:00:50|   Butterfinger mklk|      false|573bbec2cd03c9af2...|         |   Butterfinger mklk|   butterfinger mklk|[butterfinger, mklk]|          null|       null|     null|  null|  unclassified|\n",
            "| 1189870|  313842|         payment|2015-04-19 14:50:10|               Drink|      false|55335e32ca817957f...|         |               Drink|               drink|             [drink]|          Food|      drink|     null|  null|          Food|\n",
            "| 2418629| 1597426|         payment|2016-05-31 01:49:16|   Add it to the tab|      false|574c8b2ccd03c9af2...|         |   Add it to the tab|   add it to the tab|[add, it, to, the...|          Food|        tab|     null|  null|          Food|\n",
            "| 2536930| 3989960|          charge|2015-06-01 08:52:22|Chipotle mouthgas...|      false|556bbad6b651bb1cf...|       😛| Chipotle mouthgasm | chipotle mouthgasm |[chipotle, mouthg...|          Food|   chipotle|   People|    😛|          Food|\n",
            "|  265062|  257022|          charge|2015-04-23 19:49:39|Death Star episod...|      false|5538ea63efb06d213...|         |Death Star episod...|death star episod...|[death, star, epi...|IllegalSarcasm|      death|     null|  null|IllegalSarcasm|\n",
            "| 3371047| 4735051|         payment|2015-06-23 03:41:08|                   $|      false|558872e51a3b58178...|         |                   $|                   $|                 [$]|          null|       null|     null|  null|  unclassified|\n",
            "| 1641294| 2421135|         payment|2015-04-21 13:16:12|              🎉🎉🎉|      false|5535eb2c07f81c6d5...|   🎉🎉🎉|                    |                    |                  []|          null|       null| Activity|    🎉|      Activity|\n",
            "| 4892258| 5697557|         payment|2016-04-30 03:25:17|🍕 Mary had a lit...|      false|5723c32dcd03c9af2...|       🍕| Mary had a littl...| mary had a littl...|[, mary, had, a, ...|          Food|       lamb|     Food|    🍕|          Food|\n",
            "| 1814561| 4158566|         payment|2016-03-14 05:35:53|                  🕶|      false|56e5eb49cd03c9af2...|       🕶|                    |                    |                  []|          null|       null|   People|    🕶|        People|\n",
            "| 3118787| 3259892|         payment|2015-06-24 22:52:07|               Bills|      false|558ad227d4aeb62b4...|         |               Bills|               bills|             [bills]|       Utility|      bills|     null|  null|       Utility|\n",
            "|  314599|  654014|         payment|2014-02-01 05:06:38|              Square|      false|52ec1060c62cecac3...|         |              Square|              square|            [square]|          null|       null|     null|  null|  unclassified|\n",
            "| 2705822| 3748522|          charge|2015-04-24 08:52:32|  Date night 3😍😍😍|      false|5539a1e0e85574138...|   😍😍😍|        Date night 3|        date night 3|    [date, night, 3]|      Activity|      night|     null|  null|      Activity|\n",
            "| 2705822| 3748522|          charge|2015-04-24 08:52:32|  Date night 3😍😍😍|      false|5539a1e0e85574138...|   😍😍😍|        Date night 3|        date night 3|    [date, night, 3]|      Activity|       date|     null|  null|      Activity|\n",
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+\n",
            "only showing top 50 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pyspark.sql.functions as F\n",
        "\n",
        "joined_df = joined_df.withColumn(\n",
        "    'emojionly',\n",
        "    F.when(F.col('value').isNull() & F.col('value2').isNotNull(), 1).otherwise(0)\n",
        ")"
      ],
      "metadata": {
        "id": "7e8QzOieMrs8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "joined_df.show(50)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3n42Wp5RMrwF",
        "outputId": "de0303c9-16d9-47c0-f18b-00132dc1bbd8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+\n",
            "|   user1|   user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|              tokens|      category|      value|category2|value2|category_class|emojionly|\n",
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+\n",
            "|  730075|  804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|        0|\n",
            "| 4682257| 1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|   [, gypsy, things]|          null|       null|     null|  null|  unclassified|        0|\n",
            "|  241386| 2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|         [furniture]|       Utility|  furniture|     null|  null|       Utility|        0|\n",
            "| 3393501| 1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|         [bleachers]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 1864528|  437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|[the, decline, of...|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 5670003|  618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|            [crepes]|          null|       null|     null|  null|  unclassified|        0|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|    Event|    🎂|         Event|        1|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍳|          Food|        1|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍰|          Food|        1|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|     ☕|          Food|        1|\n",
            "|  884214| 1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍴|          Food|        1|\n",
            "| 1168366|  735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|[shipping, and, h...|Transportation|   shipping|     null|  null|Transportation|        0|\n",
            "| 1751344| 2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning| [5/27, drycleaning]|       Utility|drycleaning|     null|  null|       Utility|        0|\n",
            "| 3658258| 3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|            [closet]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 4828390| 4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|        [u, trippin]|          null|       null|     null|  null|  unclassified|        0|\n",
            "|  179833|  942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|      foodz|     null|  null|          Food|        0|\n",
            "|  179833|  942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|         za|     null|  null|          Food|        0|\n",
            "| 8894083| 2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|[outside, lands, ...|          null|       null|     null|  null|  unclassified|        0|\n",
            "|  411633| 4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|         [groceries]|          Food|  groceries|     null|  null|          Food|        0|\n",
            "|  303413| 3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|             [chick]|          Food|      chick|     null|  null|          Food|        0|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍦|          Food|        0|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍩|          Food|        0|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|     Food|    🍪|          Food|        0|\n",
            "| 1158024| 3438972|         payment|2015-08-17 09:11:48|👯🍦🍪🍩 getting ...|      false|55d142e42b64f704e...| 👯🍦🍪🍩| getting cookie d...| getting cookie d...|[, getting, cooki...|          Food|     cookie|   People|    👯|          Food|        0|\n",
            "| 6638728| 8528091|         payment|2016-02-23 10:26:27|      Parade day🍻🍻|      false|56cbc353cd03c9af2...|     🍻🍻|          Parade day|          parade day|       [parade, day]|      Activity|        day|     Food|    🍻|      Activity|        0|\n",
            "| 3204155| 4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|[extra, cuz, i, l...|        People|       love|     null|  null|        People|        0|\n",
            "| 3204155| 4470067|         payment|2016-08-18 22:56:42|Extra cuz i love you|      false|57b5daba23e064eac...|         |Extra cuz i love you|extra cuz i love you|[extra, cuz, i, l...|          Cash|      extra|     null|  null|          Cash|        0|\n",
            "| 7104084| 7006896|         payment|2015-12-09 05:15:04|                Food|      false|56674858cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|        0|\n",
            "| 2674970| 3691801|         payment|2015-06-19 06:05:46|               Foodz|      false|55834ecb7b9372746...|         |               Foodz|               foodz|             [foodz]|          Food|      foodz|     null|  null|          Food|        0|\n",
            "| 6523579| 1323619|         payment|2016-07-04 11:16:16|             Ran old|      false|5779e31023e064eac...|         |             Ran old|             ran old|          [ran, old]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 3413708| 1589735|          charge|2015-02-05 08:02:14|          Trash bill|      false|54d2b30791bd05aa9...|         |          Trash bill|          trash bill|       [trash, bill]|       Utility|       bill|     null|  null|       Utility|        0|\n",
            "| 3413708| 1589735|          charge|2015-02-05 08:02:14|          Trash bill|      false|54d2b30791bd05aa9...|         |          Trash bill|          trash bill|       [trash, bill]|       Utility|      trash|     null|  null|       Utility|        0|\n",
            "|11339258| 1655285|          charge|2016-09-23 00:14:44|        October wifi|      false|57e4118423e064eac...|         |        October wifi|        october wifi|     [october, wifi]|       Utility|       wifi|     null|  null|       Utility|        0|\n",
            "| 4226498| 4708480|         payment|2016-04-18 01:21:10|                  🍺|      false|5713d416cd03c9af2...|       🍺|                    |                    |                  []|          null|       null|     Food|    🍺|          Food|        1|\n",
            "|  230898|  731601|         payment|2016-07-01 00:58:39|Reverse engineeri...|      false|57755dcf23e064eac...|         |Reverse engineeri...|reverse engineeri...|[reverse, enginee...|          null|       null|     null|  null|  unclassified|        0|\n",
            "|  279563|  767893|         payment|2015-05-27 07:17:22|                Time|      false|55650d12d4aeb63c2...|         |                Time|                time|              [time]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 3562883| 7420992|         payment|2016-09-29 00:16:21|     Daddy's wealthy|      false|57ebfae523e064eac...|         |     Daddy's wealthy|     daddy's wealthy|  [daddy's, wealthy]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 6471126|10046265|          charge|2016-05-18 08:00:50|   Butterfinger mklk|      false|573bbec2cd03c9af2...|         |   Butterfinger mklk|   butterfinger mklk|[butterfinger, mklk]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 1189870|  313842|         payment|2015-04-19 14:50:10|               Drink|      false|55335e32ca817957f...|         |               Drink|               drink|             [drink]|          Food|      drink|     null|  null|          Food|        0|\n",
            "| 2418629| 1597426|         payment|2016-05-31 01:49:16|   Add it to the tab|      false|574c8b2ccd03c9af2...|         |   Add it to the tab|   add it to the tab|[add, it, to, the...|          Food|        tab|     null|  null|          Food|        0|\n",
            "| 2536930| 3989960|          charge|2015-06-01 08:52:22|Chipotle mouthgas...|      false|556bbad6b651bb1cf...|       😛| Chipotle mouthgasm | chipotle mouthgasm |[chipotle, mouthg...|          Food|   chipotle|   People|    😛|          Food|        0|\n",
            "|  265062|  257022|          charge|2015-04-23 19:49:39|Death Star episod...|      false|5538ea63efb06d213...|         |Death Star episod...|death star episod...|[death, star, epi...|IllegalSarcasm|      death|     null|  null|IllegalSarcasm|        0|\n",
            "| 3371047| 4735051|         payment|2015-06-23 03:41:08|                   $|      false|558872e51a3b58178...|         |                   $|                   $|                 [$]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 1641294| 2421135|         payment|2015-04-21 13:16:12|              🎉🎉🎉|      false|5535eb2c07f81c6d5...|   🎉🎉🎉|                    |                    |                  []|          null|       null| Activity|    🎉|      Activity|        1|\n",
            "| 4892258| 5697557|         payment|2016-04-30 03:25:17|🍕 Mary had a lit...|      false|5723c32dcd03c9af2...|       🍕| Mary had a littl...| mary had a littl...|[, mary, had, a, ...|          Food|       lamb|     Food|    🍕|          Food|        0|\n",
            "| 1814561| 4158566|         payment|2016-03-14 05:35:53|                  🕶|      false|56e5eb49cd03c9af2...|       🕶|                    |                    |                  []|          null|       null|   People|    🕶|        People|        1|\n",
            "| 3118787| 3259892|         payment|2015-06-24 22:52:07|               Bills|      false|558ad227d4aeb62b4...|         |               Bills|               bills|             [bills]|       Utility|      bills|     null|  null|       Utility|        0|\n",
            "|  314599|  654014|         payment|2014-02-01 05:06:38|              Square|      false|52ec1060c62cecac3...|         |              Square|              square|            [square]|          null|       null|     null|  null|  unclassified|        0|\n",
            "| 2705822| 3748522|          charge|2015-04-24 08:52:32|  Date night 3😍😍😍|      false|5539a1e0e85574138...|   😍😍😍|        Date night 3|        date night 3|    [date, night, 3]|      Activity|      night|     null|  null|      Activity|        0|\n",
            "| 2705822| 3748522|          charge|2015-04-24 08:52:32|  Date night 3😍😍😍|      false|5539a1e0e85574138...|   😍😍😍|        Date night 3|        date night 3|    [date, night, 3]|      Activity|       date|     null|  null|      Activity|        0|\n",
            "+--------+--------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+\n",
            "only showing top 50 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Percent of emoji only transactions"
      ],
      "metadata": {
        "id": "t-2W1tBUNRBZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pyspark.sql.functions as F\n",
        "\n",
        "# Calculate the count of emojionly values\n",
        "emojionly_count = joined_df.select(F.col(\"emojionly\")).groupBy(F.col(\"emojionly\")).agg(F.count(F.col(\"emojionly\")).alias(\"Count_emojionly\"))\n",
        "\n",
        "# Calculate the total count\n",
        "total_count = joined_df.count()\n",
        "\n",
        "# Calculate the percentage\n",
        "emojionly_percentage = emojionly_count.withColumn(\"Percentage\", (F.col(\"Count_emojionly\") / total_count) * 100)\n",
        "\n",
        "# Show the result\n",
        "emojionly_percentage.show()\n",
        "\n",
        "#+---------+---------------+------------------+\n",
        "#|emojionly|Count_emojionly|        Percentage|\n",
        "#+---------+---------------+------------------+\n",
        "#|        1|         256023|29.054266451351523|\n",
        "#|        0|         625166| 70.94573354864848|\n",
        "#+---------+---------------+------------------+\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dDXPHAi7NRpQ",
        "outputId": "fdfbe6d7-602d-4c04-928b-6218d4318a70"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---------+---------------+------------------+\n",
            "|emojionly|Count_emojionly|        Percentage|\n",
            "+---------+---------------+------------------+\n",
            "|        1|         256023|29.054266451351523|\n",
            "|        0|         625166| 70.94573354864848|\n",
            "+---------+---------------+------------------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Top 5 most popular emoji"
      ],
      "metadata": {
        "id": "5uZ4qhktNa1H"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pyspark.sql.functions as F\n",
        "\n",
        "# Filter out null values in value2 column\n",
        "filtered_df = joined_df.filter(F.col(\"value2\").isNotNull())\n",
        "\n",
        "# Group by emoji value and count occurrences\n",
        "emoji_counts = filtered_df.groupBy(\"value2\").count()\n",
        "\n",
        "# Sort the counts in descending order\n",
        "sorted_emoji_counts = emoji_counts.orderBy(F.desc(\"count\"))\n",
        "\n",
        "# Select the top 5 emojis\n",
        "top_5_emojis = sorted_emoji_counts.select(\"value2\", \"count\").limit(5)\n",
        "\n",
        "# Show the top 5 emojis\n",
        "top_5_emojis.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zUDWEUIKNbfW",
        "outputId": "10032845-9898-442f-f42d-0d6c9f054866"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|value2|count|\n",
            "+------+-----+\n",
            "|    🍕|10969|\n",
            "|    🍻|10815|\n",
            "|    🎉| 8006|\n",
            "|    🍷| 7286|\n",
            "|    🍺| 6446|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Top 3 popular emoji Categories"
      ],
      "metadata": {
        "id": "Hgc-YHWENfUU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pyspark.sql.functions as F\n",
        "\n",
        "# Filter out null values in value2 column\n",
        "filtered_df = joined_df.filter(F.col(\"category2\").isNotNull())\n",
        "\n",
        "# Group by emoji value and count occurrences\n",
        "emoji_counts = filtered_df.groupBy(\"category2\").count()\n",
        "\n",
        "# Sort the counts in descending order\n",
        "sorted_emoji_counts = emoji_counts.orderBy(F.desc(\"count\"))\n",
        "\n",
        "top_3_emojis_cat = sorted_emoji_counts.select(\"category2\", \"count\").limit(3)\n",
        "\n",
        "top_3_emojis_cat.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oNANqm_xNepH",
        "outputId": "b225d339-8b2a-41a0-ef6e-1e60b487d02f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---------+------+\n",
            "|category2| count|\n",
            "+---------+------+\n",
            "|     Food|124945|\n",
            "|   People| 78673|\n",
            "| Activity| 33728|\n",
            "+---------+------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###  Variables to Categorise Spending Bhavior Profile\n",
        "For example, if a user has made 10 transactions, where 5 of them are food and the other 5 are activity, then the user’s spending profile will be 50% food and 50% activity."
      ],
      "metadata": {
        "id": "aJo8kxwBNmgP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import when, col\n",
        "\n",
        "\n",
        "joined_df = joined_df.withColumn(\"food\", when((col(\"category\") == \"Food\") | (col(\"category2\") == \"Food\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"people\", when((col(\"category\") == \"People\") | (col(\"category2\") == \"People\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"event\", when((col(\"category\") == \"Event\") | (col(\"category2\") == \"Event\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"activity\", when((col(\"category\") == \"Activity\") | (col(\"category2\") == \"Activity\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"travel\", when((col(\"category\") == \"Travel\") | (col(\"category2\") == \"Travel\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"transportation\", when((col(\"category\") == \"Transportation\") | (col(\"category2\") == \"Transportation\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"utility\", when((col(\"category\") == \"Utility\") | (col(\"category2\") == \"Utility\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"cash\", when((col(\"category\") == \"Cash\") | (col(\"category2\") == \"Cash\"), 1).otherwise(0))\n",
        "joined_df = joined_df.withColumn(\"illegalSarcasm\", when((col(\"category\") == \"IllegalSarcasm\") | (col(\"category2\") == \"IllegalSarcasm\"), 1).otherwise(0))\n",
        "\n",
        "joined_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "txSRVeQ1Nnb8",
        "outputId": "33b39645-6b4c-4022-dc85-998a82822fd9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|              tokens|      category|      value|category2|value2|category_class|emojionly|food|people|event|activity|travel|transportation|utility|cash|illegalSarcasm|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "| 730075| 804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|4682257|1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|   [, gypsy, things]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 241386|2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|         [furniture]|       Utility|  furniture|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|\n",
            "|3393501|1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|         [bleachers]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|1864528| 437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|[the, decline, of...|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|5670003| 618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|            [crepes]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|    Event|    🎂|         Event|        1|   0|     0|    1|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍳|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍰|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|     ☕|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍴|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|1168366| 735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|[shipping, and, h...|Transportation|   shipping|     null|  null|Transportation|        0|   0|     0|    0|       0|     0|             1|      0|   0|             0|\n",
            "|1751344|2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning| [5/27, drycleaning]|       Utility|drycleaning|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|\n",
            "|3658258|3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|            [closet]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|4828390|4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|        [u, trippin]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|      foodz|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|         za|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|8894083|2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|[outside, lands, ...|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 411633|4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|         [groceries]|          Food|  groceries|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 303413|3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|             [chick]|          Food|      chick|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "selected_joined_df = joined_df.select(\"user1\", \"food\", \"people\", \"event\", \"activity\", \"travel\", \"transportation\", \"utility\", \"cash\", \"illegalSarcasm\")\n",
        "selected_joined_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xhEVt2i7Nrd1",
        "outputId": "4bc5f2ea-e044-4c49-bba0-8e6c9dd0ceeb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "|  user1|food|people|event|activity|travel|transportation|utility|cash|illegalSarcasm|\n",
            "+-------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "| 730075|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|4682257|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 241386|   0|     0|    0|       0|     0|             0|      1|   0|             0|\n",
            "|3393501|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|1864528|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|5670003|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|   0|     0|    1|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 884214|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|1168366|   0|     0|    0|       0|     0|             1|      0|   0|             0|\n",
            "|1751344|   0|     0|    0|       0|     0|             0|      1|   0|             0|\n",
            "|3658258|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|4828390|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 179833|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 179833|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "|8894083|   0|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 411633|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "| 303413|   1|     0|    0|       0|     0|             0|      0|   0|             0|\n",
            "+-------+----+------+-----+--------+------+--------------+-------+----+--------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import functions as F\n",
        "\n",
        "# Define the columns to sum\n",
        "columns_to_sum = ['people', 'food', 'event', 'activity', 'travel', 'transportation', 'utility', 'cash', 'illegalSarcasm']\n",
        "\n",
        "# Group by 'user1' and sum the remaining columns\n",
        "summed_df = selected_joined_df.groupBy('user1').agg(\n",
        "    *[F.sum(col_name).alias(f\"total_{col_name}\") for col_name in columns_to_sum]\n",
        ")\n",
        "\n",
        "# Show the resulting DataFrame\n",
        "summed_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5mcxplmBNtoB",
        "outputId": "12598955-0130-405b-a643-42dedb500e57"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+------------+----------+-----------+--------------+------------+--------------------+-------------+----------+--------------------+\n",
            "|  user1|total_people|total_food|total_event|total_activity|total_travel|total_transportation|total_utility|total_cash|total_illegalSarcasm|\n",
            "+-------+------------+----------+-----------+--------------+------------+--------------------+-------------+----------+--------------------+\n",
            "|3809372|           0|         0|          0|             0|           0|                   0|            1|         0|                   0|\n",
            "|1862793|           0|         0|          0|             0|           0|                   0|            0|         0|                   1|\n",
            "|4445675|          10|         0|          0|             1|           0|                   5|            0|         0|                   5|\n",
            "|7466454|           1|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "| 835106|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "| 883701|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "|1755286|           2|         0|          0|             1|           0|                   0|            1|         0|                   0|\n",
            "| 243356|           0|         0|          0|             2|           0|                   0|            0|         0|                   0|\n",
            "|5638543|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "|4513456|           0|         0|          0|             2|           0|                   0|            1|         0|                   0|\n",
            "|1367193|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "| 445633|           1|         2|          0|             0|           0|                   2|            0|         0|                   0|\n",
            "|1509497|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "|2986702|           0|         0|          0|             0|           0|                   0|            1|         0|                   0|\n",
            "|2524607|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "|1739972|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "|5601528|           0|         0|          0|             1|           0|                   0|            0|         0|                   0|\n",
            "| 187647|           0|         0|          0|             1|           0|                   0|            1|         0|                   0|\n",
            "|1446447|           0|         0|          0|             0|           0|                   1|            0|         0|                   0|\n",
            "|6860733|           0|         0|          0|             0|           0|                   0|            0|         0|                   0|\n",
            "+-------+------------+----------+-----------+--------------+------------+--------------------+-------------+----------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Reshaping the data:\n",
        "from pyspark.sql.functions import array, col, explode, lit, struct\n",
        "from pyspark.sql import DataFrame\n",
        "from typing import Iterable\n",
        "\n",
        "\n",
        "# Reshaping the data:\n",
        "def melt(\n",
        "        df: DataFrame,\n",
        "        id_vars: Iterable[str], value_vars: Iterable[str],\n",
        "        var_name: str=\"variable\", value_name: str=\"value\") -> DataFrame:\n",
        "    \"\"\"Convert :class:`DataFrame` from wide to long format.\"\"\"\n",
        "\n",
        "    # Create array<struct<variable: str, value: ...>>\n",
        "    _vars_and_vals = array(*(\n",
        "        struct(lit(c).alias(var_name), col(c).alias(value_name))\n",
        "        for c in value_vars))\n",
        "\n",
        "    # Add to the DataFrame and explode\n",
        "    _tmp = df.withColumn(\"_vars_and_vals\", explode(_vars_and_vals))\n",
        "\n",
        "    cols = id_vars + [\n",
        "            col(\"_vars_and_vals\")[x].alias(x) for x in [var_name, value_name]]\n",
        "    return _tmp.select(*cols)"
      ],
      "metadata": {
        "id": "_L_SKRvANvxM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "melt(summed_df, id_vars=['user1'],\n",
        "            value_vars=['total_people', 'total_food', 'total_event', 'total_activity', 'total_travel', 'total_transportation', 'total_utility', 'total_cash', 'total_illegalSarcasm']).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6xDN7wJeNyU0",
        "outputId": "84188585-df2a-4adb-fb07-f3cbf44686b5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+--------------------+-----+\n",
            "|  user1|            variable|value|\n",
            "+-------+--------------------+-----+\n",
            "|3809372|        total_people|    0|\n",
            "|3809372|          total_food|    0|\n",
            "|3809372|         total_event|    0|\n",
            "|3809372|      total_activity|    0|\n",
            "|3809372|        total_travel|    0|\n",
            "|3809372|total_transportation|    0|\n",
            "|3809372|       total_utility|    1|\n",
            "|3809372|          total_cash|    0|\n",
            "|3809372|total_illegalSarcasm|    0|\n",
            "|1862793|        total_people|    0|\n",
            "|1862793|          total_food|    0|\n",
            "|1862793|         total_event|    0|\n",
            "|1862793|      total_activity|    0|\n",
            "|1862793|        total_travel|    0|\n",
            "|1862793|total_transportation|    0|\n",
            "|1862793|       total_utility|    0|\n",
            "|1862793|          total_cash|    0|\n",
            "|1862793|total_illegalSarcasm|    1|\n",
            "|4445675|        total_people|   10|\n",
            "|4445675|          total_food|    0|\n",
            "+-------+--------------------+-----+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "user_melt_df = melt(summed_df, id_vars=['user1'],\n",
        "            value_vars=['total_people', 'total_food', 'total_event', 'total_activity', 'total_travel', 'total_transportation', 'total_utility', 'total_cash', 'total_illegalSarcasm'])"
      ],
      "metadata": {
        "id": "4nvLn2JBN0p3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import sum, col, when\n",
        "\n",
        "# Create a Spark session\n",
        "spark = SparkSession.builder.getOrCreate()\n",
        "\n",
        "# Calculate the total for each user1 and variable\n",
        "total_df = user_melt_df.groupby(\"user1\", \"variable\").agg(sum(\"value\").alias(\"total\"))\n",
        "\n",
        "# Join the total_df back to the user_melt_df\n",
        "joined_df2 = user_melt_df.join(total_df, [\"user1\", \"variable\"], \"left\")\n",
        "\n",
        "# Replace null values with 0\n",
        "joined_df2 = joined_df2.fillna(0, subset=[\"total\"])\n",
        "\n",
        "# Calculate the percentage column\n",
        "percentage_df = joined_df2.withColumn(\"percentage\", when(col(\"total\") == 0, 0).otherwise(col(\"value\") / col(\"total\") * 100))\n",
        "\n",
        "# Sort by user1\n",
        "sorted_df = percentage_df.orderBy(col(\"user1\").desc())\n",
        "\n",
        "# Show the resulting DataFrame\n",
        "sorted_df.show(50)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x2Hz8ygHN2nJ",
        "outputId": "c4e9737a-efac-4a3d-b2a1-4f6b3ddd77bf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+--------------------+-----+-----+----------+\n",
            "|   user1|            variable|value|total|percentage|\n",
            "+--------+--------------------+-----+-----+----------+\n",
            "|15930503|         total_event|    0|    0|       0.0|\n",
            "|15930503|      total_activity|    0|    0|       0.0|\n",
            "|15930503|total_illegalSarcasm|    0|    0|       0.0|\n",
            "|15930503|          total_cash|    0|    0|       0.0|\n",
            "|15930503|        total_people|    0|    0|       0.0|\n",
            "|15930503|          total_food|    0|    0|       0.0|\n",
            "|15930503|       total_utility|    0|    0|       0.0|\n",
            "|15930503|total_transportation|    0|    0|       0.0|\n",
            "|15930503|        total_travel|    0|    0|       0.0|\n",
            "|15908361|          total_cash|    0|    0|       0.0|\n",
            "|15908361|      total_activity|    0|    0|       0.0|\n",
            "|15908361|          total_food|    0|    0|       0.0|\n",
            "|15908361|         total_event|    1|    1|     100.0|\n",
            "|15908361|total_illegalSarcasm|    0|    0|       0.0|\n",
            "|15908361|        total_people|    0|    0|       0.0|\n",
            "|15908361|       total_utility|    0|    0|       0.0|\n",
            "|15908361|total_transportation|    0|    0|       0.0|\n",
            "|15908361|        total_travel|    0|    0|       0.0|\n",
            "|15900347|          total_food|    0|    0|       0.0|\n",
            "|15900347|      total_activity|    0|    0|       0.0|\n",
            "|15900347|total_illegalSarcasm|    0|    0|       0.0|\n",
            "|15900347|          total_cash|    0|    0|       0.0|\n",
            "|15900347|        total_people|    0|    0|       0.0|\n",
            "|15900347|         total_event|    0|    0|       0.0|\n",
            "|15900347|        total_travel|    0|    0|       0.0|\n",
            "|15900347|total_transportation|    0|    0|       0.0|\n",
            "|15900347|       total_utility|    0|    0|       0.0|\n",
            "|15841342|         total_event|    0|    0|       0.0|\n",
            "|15841342|      total_activity|    0|    0|       0.0|\n",
            "|15841342|        total_people|    0|    0|       0.0|\n",
            "|15841342|          total_cash|    0|    0|       0.0|\n",
            "|15841342|        total_travel|    0|    0|       0.0|\n",
            "|15841342|          total_food|    0|    0|       0.0|\n",
            "|15841342|total_illegalSarcasm|    0|    0|       0.0|\n",
            "|15841342|total_transportation|    0|    0|       0.0|\n",
            "|15841342|       total_utility|    0|    0|       0.0|\n",
            "|15792645|         total_event|    0|    0|       0.0|\n",
            "|15792645|          total_cash|    0|    0|       0.0|\n",
            "|15792645|      total_activity|    1|    1|     100.0|\n",
            "|15792645|total_transportation|    0|    0|       0.0|\n",
            "|15792645|          total_food|    0|    0|       0.0|\n",
            "|15792645|total_illegalSarcasm|    0|    0|       0.0|\n",
            "|15792645|        total_people|    0|    0|       0.0|\n",
            "|15792645|        total_travel|    0|    0|       0.0|\n",
            "|15792645|       total_utility|    0|    0|       0.0|\n",
            "|15786828|      total_activity|    0|    0|       0.0|\n",
            "|15786828|          total_cash|    0|    0|       0.0|\n",
            "|15786828|        total_people|    0|    0|       0.0|\n",
            "|15786828|         total_event|    0|    0|       0.0|\n",
            "|15786828|        total_travel|    0|    0|       0.0|\n",
            "+--------+--------------------+-----+-----+----------+\n",
            "only showing top 50 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Dynamic Spending Profile.\n",
        "\n",
        "Let’s explore how a user’s spending profile is evolving over her lifetime in Venmo."
      ],
      "metadata": {
        "id": "vZI40Xx6N5Qc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "joined_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p-Ovl81sN5xY",
        "outputId": "953e730a-908e-4db2-85bf-024409c37da0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|   emojis|           text_only|     text_only_lower|              tokens|      category|      value|category2|value2|category_class|emojionly|food|people|event|activity|travel|transportation|utility|cash|illegalSarcasm|year_and_month|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+\n",
            "| 730075| 804466|         payment|2016-05-26 04:46:45|                Food|      false|57461d46cd03c9af2...|         |                Food|                food|              [food]|          Food|       food|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-05|\n",
            "|4682257|1870271|         payment|2016-02-24 09:14:12|     🔮 gypsy things|      false|56cd03e4cd03c9af2...|       🔮|        gypsy things|        gypsy things|   [, gypsy, things]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2016-02|\n",
            "| 241386|2580543|         payment|2015-05-17 06:00:19|           Furniture|      false|5557cc0407f81c33e...|         |           Furniture|           furniture|         [furniture]|       Utility|  furniture|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|       2015-05|\n",
            "|3393501|1791591|         payment|2015-03-14 00:24:53|           Bleachers|      false|55031d67cd03c9af2...|         |           Bleachers|           bleachers|         [bleachers]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2015-03|\n",
            "|1864528| 437516|         payment|2015-12-07 00:51:11|The decline of we...|      false|5664677fcd03c9af2...|         |The decline of we...|the decline of we...|[the, decline, of...|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2015-12|\n",
            "|5670003| 618376|         payment|2016-07-17 03:50:26|              Crepes|      false|578a9e1223e064eac...|         |              Crepes|              crepes|            [crepes]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2016-07|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|    Event|    🎂|         Event|        1|   0|     0|    1|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍳|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍰|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|     ☕|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "| 884214|1386238|         payment|2016-04-03 23:50:03|          ☕️🍴🍰🍳🎂|      false|570149bbcd03c9af2...|☕🍴🍰🍳🎂|                   ️|                   ️|                 [️]|          null|       null|     Food|    🍴|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "|1168366| 735212|         payment|2014-08-18 22:51:21|Shipping and hand...|      false|53f220fa7d0b0354e...|         |Shipping and hand...|shipping and hand...|[shipping, and, h...|Transportation|   shipping|     null|  null|Transportation|        0|   0|     0|    0|       0|     0|             1|      0|   0|             0|       2014-08|\n",
            "|1751344|2105340|          charge|2015-05-29 05:17:51|    5/27 drycleaning|      false|5567940fe85574434...|         |    5/27 drycleaning|    5/27 drycleaning| [5/27, drycleaning]|       Utility|drycleaning|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|       2015-05|\n",
            "|3658258|3655859|          charge|2015-08-28 09:48:26|              Closet|      false|55dfcbfacd03c9af2...|         |              Closet|              closet|            [closet]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2015-08|\n",
            "|4828390|4461134|         payment|2016-02-23 03:57:19|           u trippin|      false|56cb681fcd03c9af2...|         |           u trippin|           u trippin|        [u, trippin]|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2016-02|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|      foodz|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-05|\n",
            "| 179833| 942891|         payment|2015-05-13 05:23:41|            Za foodZ|      false|55527d6d1a3b58730...|         |            Za foodZ|            za foodz|         [za, foodz]|          Food|         za|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-05|\n",
            "|8894083|2400080|         payment|2016-04-01 00:28:19|  OUTSIDE LANDS 2k16|      false|56fd5e33cd03c9af2...|         |  OUTSIDE LANDS 2k16|  outside lands 2k16|[outside, lands, ...|          null|       null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2016-04|\n",
            "| 411633|4172797|          charge|2016-08-02 02:30:22|           Groceries|      false|579fa34e23e064eac...|         |           Groceries|           groceries|         [groceries]|          Food|  groceries|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-08|\n",
            "| 303413|3245733|          charge|2016-01-17 02:31:25|               Chick|      false|569a8c7dcd03c9af2...|         |               Chick|               chick|             [chick]|          Food|      chick|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2016-01|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+---------+--------------------+--------------------+--------------------+--------------+-----------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "joined_df = joined_df.withColumn(\"year_and_month\", concat(year(col(\"datetime\")).cast(StringType() ), lit(\"-\"),\n",
        "                                              when(month(col(\"datetime\")) <10,\n",
        "                                                   concat(lit(\"0\"), month(col(\"datetime\")).cast(StringType()))).otherwise(month(col(\"datetime\")).cast(StringType())) ))"
      ],
      "metadata": {
        "id": "pTJrFx7KN8Ti"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.window import Window\n",
        "window_val = (Window.partitionBy('user1','category_class').orderBy('year_and_month')\n",
        "             .rowsBetween(Window.unboundedPreceding, 0))\n",
        "window_totals = (Window.partitionBy('user1','year_and_month')\n",
        "             .rowsBetween(Window.unboundedPreceding, 0))"
      ],
      "metadata": {
        "id": "f-Vx01cfN9rd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.window import Window\n",
        "from pyspark.sql.functions import count, col, sum, row_number\n",
        "\n",
        "# Define the window specification for dynamic calculations\n",
        "window_val = Window.partitionBy('user1', 'year_and_month').orderBy('category_class')\n",
        "window_totals = Window.partitionBy('user1', 'year_and_month')\n",
        "\n",
        "# Calculate the dynamic spending Venmo\n",
        "dynamic = joined_df.filter(col(\"category_class\").isNotNull()).groupBy(\"user1\", \"year_and_month\", \"category_class\") \\\n",
        "    .agg(count(\"category_class\").alias(\"transactions\"))\n",
        "\n",
        "dynamic_spending_venmo = dynamic.withColumn(\"total_transactions_category\", sum(col(\"transactions\")).over(window_val)) \\\n",
        "    .withColumn(\"total_transactions_month\", sum(col(\"transactions\")).over(window_totals \\\n",
        "        .rowsBetween(Window.unboundedPreceding, Window.unboundedFollowing))) \\\n",
        "    .withColumn(\"dynamic_percentage\", 100 * (col(\"transactions\") / col(\"total_transactions_month\"))) \\\n",
        "    .withColumn(\"month_count\", dense_rank().over(Window.partitionBy(\"user1\").orderBy(col(\"year_and_month\").desc())) - 1)\n",
        "\n",
        "# Show the resulting DataFrame\n",
        "dynamic_spending_venmo_filtered = dynamic_spending_venmo.filter((col(\"month_count\") >= 0) & (col(\"month_count\") <= 12))\n",
        "dynamic_spending_venmo_filtered.orderBy(col(\"month_count\").desc()).show(70)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MnR9q0ZlOCi_",
        "outputId": "412eafb2-cc41-4473-c006-7d691f7105d2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+--------------+--------------+------------+---------------------------+------------------------+------------------+-----------+\n",
            "|  user1|year_and_month|category_class|transactions|total_transactions_category|total_transactions_month|dynamic_percentage|month_count|\n",
            "+-------+--------------+--------------+------------+---------------------------+------------------------+------------------+-----------+\n",
            "| 820531|       2015-11|      Activity|           5|                          5|                      10|              50.0|         12|\n",
            "| 820531|       2015-11|       Utility|           5|                         10|                      10|              50.0|         12|\n",
            "| 148281|       2013-09|          Food|           1|                          1|                       1|             100.0|         11|\n",
            "| 820531|       2015-12|      Activity|           6|                          6|                      12|              50.0|         11|\n",
            "| 820531|       2015-12|       Utility|           6|                         12|                      12|              50.0|         11|\n",
            "| 148281|       2014-06|  unclassified|           1|                          1|                       1|             100.0|         10|\n",
            "| 820531|       2016-01|      Activity|           7|                          7|                      14|              50.0|         10|\n",
            "| 476852|       2014-01|  unclassified|           1|                          1|                       1|             100.0|         10|\n",
            "| 820531|       2016-01|       Utility|           7|                         14|                      14|              50.0|         10|\n",
            "|1246259|       2014-12|          Food|           1|                          1|                       1|             100.0|         10|\n",
            "|1489625|       2014-11|  unclassified|           1|                          1|                       1|             100.0|         10|\n",
            "|  77138|       2012-10|  unclassified|           1|                          1|                       1|             100.0|          9|\n",
            "| 148281|       2014-09|  unclassified|           1|                          1|                       1|             100.0|          9|\n",
            "| 610084|       2014-03|          Food|           1|                          1|                       1|             100.0|          9|\n",
            "| 476852|       2014-04|          Food|           1|                          1|                       1|             100.0|          9|\n",
            "| 820531|       2016-02|      Activity|           6|                          6|                      12|              50.0|          9|\n",
            "|1246259|       2015-02|  unclassified|           1|                          1|                       1|             100.0|          9|\n",
            "| 820531|       2016-02|       Utility|           6|                         12|                      12|              50.0|          9|\n",
            "|1489625|       2015-03|  unclassified|           1|                          1|                       1|             100.0|          9|\n",
            "|4188588|       2015-11|          Cash|           1|                          1|                       1|             100.0|          9|\n",
            "| 148281|       2014-10|      Activity|           2|                          2|                       3| 66.66666666666666|          8|\n",
            "| 125369|       2014-10|  unclassified|           1|                          1|                       1|             100.0|          8|\n",
            "| 148281|       2014-10|  unclassified|           1|                          3|                       3| 33.33333333333333|          8|\n",
            "|  77138|       2012-11|      Activity|           1|                          1|                       2|              50.0|          8|\n",
            "| 175908|       2013-07|          Food|           1|                          1|                       2|              50.0|          8|\n",
            "|  77138|       2012-11|       Utility|           1|                          2|                       2|              50.0|          8|\n",
            "| 175908|       2013-07|       Utility|           1|                          2|                       2|              50.0|          8|\n",
            "| 257334|       2013-04|        People|           1|                          1|                       1|             100.0|          8|\n",
            "| 317427|       2013-10|      Activity|           1|                          1|                       2|              50.0|          8|\n",
            "| 266819|       2013-10|       Utility|           2|                          2|                       2|             100.0|          8|\n",
            "| 317427|       2013-10|          Food|           1|                          2|                       2|              50.0|          8|\n",
            "| 590439|       2014-09|Transportation|           1|                          1|                       1|             100.0|          8|\n",
            "| 339157|       2013-07|       Utility|           1|                          1|                       1|             100.0|          8|\n",
            "| 610084|       2014-05|       Utility|           1|                          1|                       2|              50.0|          8|\n",
            "| 402321|       2014-03|Transportation|           1|                          1|                       1|             100.0|          8|\n",
            "| 610084|       2014-05|  unclassified|           1|                          2|                       2|              50.0|          8|\n",
            "| 424506|       2014-03|          Food|           1|                          1|                       2|              50.0|          8|\n",
            "| 614730|       2014-03|          Food|           1|                          1|                       2|              50.0|          8|\n",
            "| 424506|       2014-03|        People|           1|                          2|                       2|              50.0|          8|\n",
            "| 614730|       2014-03|Transportation|           1|                          2|                       2|              50.0|          8|\n",
            "| 476852|       2014-05|       Utility|           1|                          1|                       1|             100.0|          8|\n",
            "| 820531|       2016-03|      Activity|           8|                          8|                      16|              50.0|          8|\n",
            "|1246259|       2015-03|      Activity|           1|                          1|                       2|              50.0|          8|\n",
            "| 820531|       2016-03|       Utility|           8|                         16|                      16|              50.0|          8|\n",
            "|1246259|       2015-03|          Food|           1|                          2|                       2|              50.0|          8|\n",
            "|2242328|       2014-10|      Activity|           1|                          1|                       2|              50.0|          8|\n",
            "|1333850|       2014-11|  unclassified|           1|                          1|                       1|             100.0|          8|\n",
            "|2242328|       2014-10|        People|           1|                          2|                       2|              50.0|          8|\n",
            "|1489625|       2015-05|  unclassified|           2|                          2|                       2|             100.0|          8|\n",
            "|4188588|       2016-01|          Cash|           2|                          2|                       2|             100.0|          8|\n",
            "| 283863|       2014-01|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 249401|       2013-09|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 148281|       2014-12|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 125369|       2014-12|      Activity|           1|                          1|                       1|             100.0|          7|\n",
            "| 163866|       2014-05|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 141885|       2013-04|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 175908|       2013-10|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 150496|       2013-05|Transportation|           1|                          1|                       1|             100.0|          7|\n",
            "| 180413|       2013-04|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 152261|       2013-11|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 190215|       2014-03|          Food|           1|                          1|                       1|             100.0|          7|\n",
            "| 180412|       2014-01|          Food|           2|                          2|                       2|             100.0|          7|\n",
            "| 271099|       2014-02|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 190382|       2012-12|        Travel|           1|                          1|                       1|             100.0|          7|\n",
            "| 317427|       2013-12|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 244876|       2014-02|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 339157|       2013-11|         Event|           1|                          1|                       3| 33.33333333333333|          7|\n",
            "| 165455|       2013-02|  unclassified|           1|                          1|                       1|             100.0|          7|\n",
            "| 339157|       2013-11|          Food|           1|                          2|                       3| 33.33333333333333|          7|\n",
            "| 257334|       2013-06|Transportation|           1|                          1|                       1|             100.0|          7|\n",
            "+-------+--------------+--------------+------------+---------------------------+------------------------+------------------+-----------+\n",
            "only showing top 70 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dynamic_spending_venmo_filtered2 = dynamic_spending_venmo_filtered.select(\"user1\",\"category_class\",\"dynamic_percentage\",\"month_count\")"
      ],
      "metadata": {
        "id": "uRkFhsRHdKOz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "dynamic_spending_venmo_filtered2.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cwl6uBMgfOoT",
        "outputId": "12cd5cfc-9e4d-4785-ad70-64bdcf85160c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+------------------+-----------+\n",
            "|user1|category_class|dynamic_percentage|month_count|\n",
            "+-----+--------------+------------------+-----------+\n",
            "|    3|       Utility|             100.0|          0|\n",
            "|   13|  unclassified|             100.0|          0|\n",
            "|   19|      Activity|             100.0|          0|\n",
            "|   43|  unclassified|             100.0|          1|\n",
            "|   43|          Food|             100.0|          0|\n",
            "|   52|          Food|             100.0|          0|\n",
            "|  112|  unclassified|             100.0|          0|\n",
            "|  164|  unclassified|             100.0|          0|\n",
            "|  243|       Utility|             100.0|          0|\n",
            "|  347|          Food|             100.0|          0|\n",
            "|  747|Transportation|             100.0|          3|\n",
            "|  747|Transportation|             100.0|          2|\n",
            "|  747|          Food|             100.0|          1|\n",
            "|  747|          Cash|              50.0|          0|\n",
            "|  747|       Utility|              50.0|          0|\n",
            "|  830|        People|              25.0|          0|\n",
            "|  830|          Food|              75.0|          0|\n",
            "|  879|        People|             100.0|          0|\n",
            "|  906|  unclassified|             100.0|          0|\n",
            "| 1255|          Food|             100.0|          0|\n",
            "+-----+--------------+------------------+-----------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "joined_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Lp85Im38Tqh0",
        "outputId": "eaebc92a-8986-435b-d545-cff06f0de425"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+-------+----------------+-------------------+--------------------+-----------+--------------------+--------+--------------------+--------------------+--------------------+--------------+---------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+-----+----------------------+------------------+\n",
            "|user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|  emojis|           text_only|     text_only_lower|              tokens|      category|    value|category2|value2|category_class|emojionly|food|people|event|activity|travel|transportation|utility|cash|illegalSarcasm|year_and_month|month|first_transaction_date|months_since_first|\n",
            "+-----+-------+----------------+-------------------+--------------------+-----------+--------------------+--------+--------------------+--------------------+--------------------+--------------+---------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+-----+----------------------+------------------+\n",
            "|    3|7854140|         payment|2016-10-09 03:36:13|Check out this re...|      false|57f958bd23e064eac...|        |Check out this re...|check out this re...|[check, out, this...|       Utility|    check|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|       2016-10|   10|   2016-10-09 03:36:13|                 0|\n",
            "|   13|1941801|          charge|2014-10-16 22:04:27|Starfinder Winter...|      false|543fde7b7d0b0354e...|        |Starfinder Winter...|starfinder winter...|[starfinder, wint...|          null|     null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2014-10|   10|   2014-10-16 22:04:27|                 0|\n",
            "|   19|  85471|         payment|2014-03-17 21:43:39|                Pool|      false|53270a1bb19abafe2...|        |                Pool|                pool|              [pool]|      Activity|     pool|     null|  null|      Activity|        0|   0|     0|    0|       1|     0|             0|      0|   0|             0|       2014-03|    3|   2014-03-17 21:43:39|                 0|\n",
            "|   43|  49778|         payment|2013-01-08 09:39:09|          ☔ & swords|      false|50eb798d25ee44b9a...|       ☔|            & swords|            & swords|       [, &, swords]|          null|     null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2013-01|    1|   2013-01-08 09:39:09|                 0|\n",
            "|   52|    225|         payment|2012-08-06 20:48:32|    Curiosity coffee|      false|501fcb30c5b30a098...|        |    Curiosity coffee|    curiosity coffee| [curiosity, coffee]|          Food|   coffee|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2012-08|    8|   2012-08-06 20:48:32|                 0|\n",
            "|  112|     12|         payment|2015-11-05 02:39:29|for microwave rep...|      false|563a50e1cd03c9af2...|        |for microwave rep...|for microwave rep...|[for, microwave, ...|          null|     null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2015-11|   11|   2015-11-05 02:39:29|                 0|\n",
            "|  164| 486473|         payment|2016-03-13 10:48:28|              Effies|      false|56e4d4fccd03c9af2...|        |              Effies|              effies|            [effies]|          null|     null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2016-03|    3|   2016-03-13 10:48:28|                 0|\n",
            "|  243|      9|          charge|2012-12-28 08:52:01|                Maid|      false|50dcedeb25ee44b9a...|        |                Maid|                maid|              [maid]|       Utility|     maid|     null|  null|       Utility|        0|   0|     0|    0|       0|     0|             0|      1|   0|             0|       2012-12|   12|   2012-12-28 08:52:01|                 0|\n",
            "|  347|2131067|         payment|2015-05-11 10:54:34|Uncle Ben's Rice ...|      false|555027fad4aeb646b...|        |Uncle Ben's Rice ...|uncle ben's rice ...|[uncle, ben's, ri...|          Food|     rice|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-05|    5|   2015-05-11 10:54:34|                 0|\n",
            "|  747| 158173|          charge|2013-12-24 10:31:05|                 Cab|      false|52b8f1e9d56b6bac5...|        |                 Cab|                 cab|               [cab]|Transportation|      cab|     null|  null|Transportation|        0|   0|     0|    0|       0|     0|             1|      0|   0|             0|       2013-12|   12|   2013-03-18 21:38:39|                 9|\n",
            "|  747|   1009|          charge|2013-03-18 21:38:39|   Taxi oh dangggggg|      false|5147285325ee44b9a...|        |   Taxi oh dangggggg|   taxi oh dangggggg|[taxi, oh, danggg...|Transportation|     taxi|     null|  null|Transportation|        0|   0|     0|    0|       0|     0|             1|      0|   0|             0|       2013-03|    3|   2013-03-18 21:38:39|                 0|\n",
            "|  830|1291070|         payment|2015-12-13 12:28:09|        PDT 😆🍟🍸🍾|      false|566cf3dacd03c9af2...|😆🍟🍸🍾|                PDT |                pdt |               [pdt]|          null|     null|   People|    😆|        People|        1|   0|     1|    0|       0|     0|             0|      0|   0|             0|       2015-12|   12|   2015-12-13 12:28:09|                 0|\n",
            "|  830|1291070|         payment|2015-12-13 12:28:09|        PDT 😆🍟🍸🍾|      false|566cf3dacd03c9af2...|😆🍟🍸🍾|                PDT |                pdt |               [pdt]|          null|     null|     Food|    🍟|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-12|   12|   2015-12-13 12:28:09|                 0|\n",
            "|  830|1291070|         payment|2015-12-13 12:28:09|        PDT 😆🍟🍸🍾|      false|566cf3dacd03c9af2...|😆🍟🍸🍾|                PDT |                pdt |               [pdt]|          null|     null|     Food|    🍾|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-12|   12|   2015-12-13 12:28:09|                 0|\n",
            "|  830|1291070|         payment|2015-12-13 12:28:09|        PDT 😆🍟🍸🍾|      false|566cf3dacd03c9af2...|😆🍟🍸🍾|                PDT |                pdt |               [pdt]|          null|     null|     Food|    🍸|          Food|        1|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-12|   12|   2015-12-13 12:28:09|                 0|\n",
            "|  879|  44731|         payment|2013-07-12 03:42:27|for swindling the...|      false|51df155f7de518fa3...|        |for swindling the...|for swindling the...|[for, swindling, ...|        People|      guy|     null|  null|        People|        0|   0|     1|    0|       0|     0|             0|      0|   0|             0|       2013-07|    7|   2013-07-12 03:42:27|                 0|\n",
            "|  906| 126820|         payment|2013-11-29 14:17:08|              Tx! :)|      false|52983164d56b6bac5...|        |              Tx! :)|              tx! :)|           [tx!, :)]|          null|     null|     null|  null|  unclassified|        0|   0|     0|    0|       0|     0|             0|      0|   0|             0|       2013-11|   11|   2013-11-29 14:17:08|                 0|\n",
            "| 1255|5935714|         payment|2015-12-18 01:58:06|  Chocolate extreme!|      false|5672f7aecd03c9af2...|        |  Chocolate extreme!|  chocolate extreme!|[chocolate, extre...|          Food|chocolate|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2015-12|   12|   2015-12-18 01:58:06|                 0|\n",
            "| 1632| 791046|         payment|2016-03-09 00:29:45|Flowers in Houn's...|      false|56defdf9cd03c9af2...|      😭|Flowers in Houn's...|flowers in houn's...|[flowers, in, hou...|IllegalSarcasm|      rip|   People|    😭|IllegalSarcasm|        0|   0|     1|    0|       0|     0|             0|      0|   0|             1|       2016-03|    3|   2016-03-09 00:29:45|                 0|\n",
            "| 1636|   1722|         payment|2013-05-22 13:36:44|       Food and beer|      false|519c696825ee44b9a...|        |       Food and beer|       food and beer|   [food, and, beer]|          Food|     food|     null|  null|          Food|        0|   1|     0|    0|       0|     0|             0|      0|   0|             0|       2013-05|    5|   2013-05-22 13:36:44|                 0|\n",
            "+-----+-------+----------------+-------------------+--------------------+-----------+--------------------+--------+--------------------+--------------------+--------------------+--------------+---------+---------+------+--------------+---------+----+------+-----+--------+------+--------------+-------+----+--------------+--------------+-----+----------------------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_df = joined_df.select(\"user1\", \"category_class\",  \"datetime\")\n",
        "\n",
        "plot_df.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wXbe9WkNSoBB",
        "outputId": "29e190d0-582f-4f38-f78d-329260fabd70"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------------------+\n",
            "|user1|category_class|           datetime|\n",
            "+-----+--------------+-------------------+\n",
            "|    3|       Utility|2016-10-09 03:36:13|\n",
            "|   13|  unclassified|2014-10-16 22:04:27|\n",
            "|   19|      Activity|2014-03-17 21:43:39|\n",
            "|   43|  unclassified|2013-01-08 09:39:09|\n",
            "|   52|          Food|2012-08-06 20:48:32|\n",
            "|  112|  unclassified|2015-11-05 02:39:29|\n",
            "|  164|  unclassified|2016-03-13 10:48:28|\n",
            "|  243|       Utility|2012-12-28 08:52:01|\n",
            "|  347|          Food|2015-05-11 10:54:34|\n",
            "|  747|Transportation|2013-12-24 10:31:05|\n",
            "|  747|Transportation|2013-03-18 21:38:39|\n",
            "|  830|        People|2015-12-13 12:28:09|\n",
            "|  830|          Food|2015-12-13 12:28:09|\n",
            "|  830|          Food|2015-12-13 12:28:09|\n",
            "|  830|          Food|2015-12-13 12:28:09|\n",
            "|  879|        People|2013-07-12 03:42:27|\n",
            "|  906|  unclassified|2013-11-29 14:17:08|\n",
            "| 1255|          Food|2015-12-18 01:58:06|\n",
            "| 1632|IllegalSarcasm|2016-03-09 00:29:45|\n",
            "| 1636|          Food|2013-05-22 13:36:44|\n",
            "+-----+--------------+-------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.window import Window\n",
        "from pyspark.sql.functions import col, when, datediff, min\n",
        "\n",
        "plot_df = plot_df.withColumn(\n",
        "    \"first_day\", min(\"datetime\").over(Window.partitionBy(\"user\"))\n",
        ").withColumn(\n",
        "    \"life_point\",\n",
        "    when(col(\"datetime\") == col(\"first_day\"), 0)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 30, 1)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 60, 2)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 90, 3)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 120, 4)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 150, 5)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 180, 6)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 210, 7)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 240, 8)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 270, 9)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 300, 10)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 330, 11)\n",
        "    .when(datediff(col(\"datetime\"), col(\"first_day\")) <= 360, 12)\n",
        ")\n",
        "\n",
        "plot_df = plot_df.filter(col(\"life_point\") <= 12)\n",
        "\n",
        "plot_df.show(50)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e-BqbVq0TbZ2",
        "outputId": "e984e6df-a95e-4b57-e783-79bfe0a9ce4c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------------------+-------------------+----------+\n",
            "|user1|category_class|           datetime|          first_day|life_point|\n",
            "+-----+--------------+-------------------+-------------------+----------+\n",
            "|   43|  unclassified|2013-01-08 09:39:09|2012-04-12 11:40:34|        10|\n",
            "|   52|          Food|2012-08-06 20:48:32|2012-04-12 11:40:34|         4|\n",
            "|  243|       Utility|2012-12-28 08:52:01|2012-04-12 11:40:34|         9|\n",
            "|  747|Transportation|2013-03-18 21:38:39|2012-04-12 11:40:34|        12|\n",
            "| 2391|  unclassified|2013-01-23 09:16:42|2012-04-12 11:40:34|        10|\n",
            "| 2721|  unclassified|2012-05-26 00:52:22|2012-04-12 11:40:34|         2|\n",
            "| 2794|          Food|2012-07-01 09:08:55|2012-04-12 11:40:34|         3|\n",
            "| 2794|        Travel|2012-04-29 23:29:52|2012-04-12 11:40:34|         1|\n",
            "| 2794|          Food|2012-04-29 23:29:52|2012-04-12 11:40:34|         1|\n",
            "| 3310|       Utility|2013-03-04 11:48:39|2012-04-12 11:40:34|        11|\n",
            "| 3664|          Food|2012-06-04 23:42:59|2012-04-12 11:40:34|         2|\n",
            "| 3664|          Food|2012-06-04 23:42:59|2012-04-12 11:40:34|         2|\n",
            "| 4262|  unclassified|2012-10-12 06:16:42|2012-04-12 11:40:34|         7|\n",
            "| 5066|IllegalSarcasm|2013-04-04 00:27:30|2012-04-12 11:40:34|        12|\n",
            "| 6329|          Food|2013-01-03 04:05:27|2012-04-12 11:40:34|         9|\n",
            "| 6612|       Utility|2012-12-16 04:04:45|2012-04-12 11:40:34|         9|\n",
            "| 6612|       Utility|2012-12-16 04:04:45|2012-04-12 11:40:34|         9|\n",
            "|25535|          Food|2012-11-18 16:28:00|2012-04-12 11:40:34|         8|\n",
            "|25535|          Food|2012-10-14 07:09:06|2012-04-12 11:40:34|         7|\n",
            "|26384|  unclassified|2012-12-24 07:46:31|2012-04-12 11:40:34|         9|\n",
            "|29894|  unclassified|2013-03-09 07:30:36|2012-04-12 11:40:34|        12|\n",
            "|30139|       Utility|2012-10-26 07:12:24|2012-04-12 11:40:34|         7|\n",
            "|31249|  unclassified|2013-01-01 08:51:05|2012-04-12 11:40:34|         9|\n",
            "|33383|          Food|2013-02-11 06:45:36|2012-04-12 11:40:34|        11|\n",
            "|33764|  unclassified|2012-06-23 10:25:12|2012-04-12 11:40:34|         3|\n",
            "|34084|          Food|2012-07-29 22:42:21|2012-04-12 11:40:34|         4|\n",
            "|34336|  unclassified|2012-04-24 06:12:54|2012-04-12 11:40:34|         1|\n",
            "|35635|  unclassified|2012-12-29 16:30:57|2012-04-12 11:40:34|         9|\n",
            "|36491|IllegalSarcasm|2012-06-30 06:28:12|2012-04-12 11:40:34|         3|\n",
            "|36667|  unclassified|2013-01-21 01:04:11|2012-04-12 11:40:34|        10|\n",
            "|36712|        People|2012-09-08 04:44:30|2012-04-12 11:40:34|         5|\n",
            "|36712|          Food|2012-09-08 04:44:30|2012-04-12 11:40:34|         5|\n",
            "|37289|       Utility|2013-03-23 09:22:37|2012-04-12 11:40:34|        12|\n",
            "|37818|       Utility|2012-06-01 18:57:16|2012-04-12 11:40:34|         2|\n",
            "|38864|  unclassified|2013-02-21 00:02:57|2012-04-12 11:40:34|        11|\n",
            "|39108|  unclassified|2013-04-01 02:59:41|2012-04-12 11:40:34|        12|\n",
            "|39582|          Food|2013-01-05 10:49:37|2012-04-12 11:40:34|         9|\n",
            "|39678|  unclassified|2012-10-11 07:30:22|2012-04-12 11:40:34|         7|\n",
            "|39867|       Utility|2012-11-04 06:14:06|2012-04-12 11:40:34|         7|\n",
            "|39867|      Activity|2012-11-04 06:14:06|2012-04-12 11:40:34|         7|\n",
            "|39867|          Cash|2012-11-04 06:14:06|2012-04-12 11:40:34|         7|\n",
            "|41680|  unclassified|2012-05-18 03:45:28|2012-04-12 11:40:34|         2|\n",
            "|42928|      Activity|2012-10-12 09:15:12|2012-04-12 11:40:34|         7|\n",
            "|43040|  unclassified|2012-11-04 10:35:48|2012-04-12 11:40:34|         7|\n",
            "|43277|  unclassified|2012-05-07 06:34:34|2012-04-12 11:40:34|         1|\n",
            "|43994|  unclassified|2012-12-24 08:10:51|2012-04-12 11:40:34|         9|\n",
            "|44669|  unclassified|2013-01-07 11:13:06|2012-04-12 11:40:34|         9|\n",
            "|45957|          Food|2012-10-22 04:11:03|2012-04-12 11:40:34|         7|\n",
            "|46041|IllegalSarcasm|2012-11-09 13:17:38|2012-04-12 11:40:34|         8|\n",
            "|46372|  unclassified|2012-05-27 08:21:03|2012-04-12 11:40:34|         2|\n",
            "+-----+--------------+-------------------+-------------------+----------+\n",
            "only showing top 50 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import sum, when, col\n",
        "\n",
        "classification_by_time = plot_df.groupBy(col(\"user1\").alias(\"user1\"), \"life_point\").agg(\n",
        "    (sum(when(col(\"category_class\") == \"Activity\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_activity\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Food\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_food\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"People\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_people\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Event\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_event\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Travel\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_travel\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Transportation\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_transportation\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Utility\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_utility\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"Cash\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_cash\"),\n",
        "\n",
        "    (sum(when(col(\"category_class\") == \"IllegalSarcasm\", 1).otherwise(0)) /\n",
        "    sum(when(col(\"category_class\") != \"unclassified\", 1).otherwise(0))).alias(\"count_illegal\")\n",
        ")\n",
        "\n",
        "classification_by_time.show(100)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZwOJGd4KUDVS",
        "outputId": "b0003911-a2a5-4551-b358-9505a94a3e68"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----------+------------------+------------------+------------+-----------+------------------+--------------------+-------------+------------------+-------------+\n",
            "| user1|life_point|    count_activity|        count_food|count_people|count_event|      count_travel|count_transportation|count_utility|        count_cash|count_illegal|\n",
            "+------+----------+------------------+------------------+------------+-----------+------------------+--------------------+-------------+------------------+-------------+\n",
            "| 86796|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|166777|        10|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 29489|         1|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|117431|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|130871|         4|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               1.0|          0.0|\n",
            "|134892|         7|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 74053|         3|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 43794|         3|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|129624|         5|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|246998|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|130048|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|141513|         5|               0.0|               0.0|         0.0|        0.0|               0.0|                 1.0|          0.0|               0.0|          0.0|\n",
            "|196931|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|224036|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 73816|        12|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|223643|        11|               0.5|               0.0|         0.0|        0.0|               0.5|                 0.0|          0.0|               0.0|          0.0|\n",
            "|113632|         2|0.3333333333333333|               0.0|         0.0|        0.0|0.6666666666666666|                 0.0|          0.0|               0.0|          0.0|\n",
            "|117886|         4|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|136798|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|139684|        10|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|163174|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|  6612|         9|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|121590|        10|               0.5|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.5|\n",
            "|127766|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|148680|        11|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 93445|         3|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|155914|        10|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|175203|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 62066|         2|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "| 76355|         6|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|157343|         7|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|170581|        11|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|216900|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|177785|        11|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|191310|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|200758|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|220382|        11|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|238560|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|142016|         6|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|159205|        11|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|180943|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|190871|         9|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "| 60575|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|174835|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|234325|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 70058|        10|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|171577|        11|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "| 89442|        10|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|118903|        11|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|135444|         6|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|147414|         8|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|171594|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|199074|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|239525|        12|               0.0|               0.0|         0.0|        0.0|               1.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|200195|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|205915|        10|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|110666|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|112187|         5|               0.0|0.6666666666666666|         0.0|        0.0|               0.0|                 0.0|          0.0|0.3333333333333333|          0.0|\n",
            "| 49304|         9|               0.0|               0.5|         0.5|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|125059|         9|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 89917|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|109088|         6|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|175364|        12|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|205046|        12|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|124323|         6|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|198716|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 81315|        12|               0.0|               0.0|         1.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|163826|        10|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 43933|        12|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|111612|         4|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|137801|         9|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|206375|        11|               0.0|               0.0|         0.0|        0.0|               0.5|                 0.0|          0.5|               0.0|          0.0|\n",
            "| 43994|         9|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 61016|        12|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 98974|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|184621|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|212078|        10|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|220747|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|110155|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|112868|        12|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 80697|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|141971|        12|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 22413|         8|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|117886|         5|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|146507|         7|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|172383|        10|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "|250747|        12|               0.0|               0.5|         0.0|        0.5|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "| 41680|         2|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 86182|        11|               0.0|               1.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|165838|        12|               0.0|               0.0|         0.0|        0.0|               0.0|                 1.0|          0.0|               0.0|          0.0|\n",
            "| 96412|         6|               0.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          1.0|               0.0|          0.0|\n",
            "| 72464|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|134519|        10|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|154853|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|185841|        12|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|226265|        12|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 63072|         2|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "| 83450|         2|               1.0|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.0|          0.0|\n",
            "|122352|        11|              null|              null|        null|       null|              null|                null|         null|              null|         null|\n",
            "|129566|         5|               0.5|               0.0|         0.0|        0.0|               0.0|                 0.0|          0.0|               0.5|          0.0|\n",
            "+------+----------+------------------+------------------+------------+-----------+------------------+--------------------+-------------+------------------+-------------+\n",
            "only showing top 100 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import expr\n",
        "\n",
        "avg_by_time_df = classification_by_time.groupBy(\"life_point\").agg(\n",
        "    avg(\"count_activity\").alias(\"avg_activity\"),\n",
        "    avg(\"count_food\").alias(\"avg_food\"),\n",
        "    avg(\"count_people\").alias(\"avg_people\"),\n",
        "    avg(\"count_event\").alias(\"avg_event\"),\n",
        "    avg(\"count_travel\").alias(\"avg_travel\"),\n",
        "    avg(\"count_transportation\").alias(\"avg_transportation\"),\n",
        "    avg(\"count_utility\").alias(\"avg_utility\"),\n",
        "    avg(\"count_cash\").alias(\"avg_cash\"),\n",
        "    avg(\"count_illegal\").alias(\"avg_illegal\")\n",
        ")\n",
        "\n",
        "melted_columns = ['avg_activity', 'avg_food', 'avg_people', 'avg_event', 'avg_travel', 'avg_transportation',\n",
        "                  'avg_utility', 'avg_cash', 'avg_illegal']\n",
        "\n",
        "melted_columns_expr = \", \".join([\"'{0}', {0}\".format(col) for col in melted_columns])\n",
        "\n",
        "avg_by_time_df = avg_by_time_df.selectExpr(\"life_point\", \"stack({0}, {1}) as (variable, value)\".format(\n",
        "    len(melted_columns), melted_columns_expr\n",
        "))\n",
        "\n",
        "avg_by_time_df.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "URKG7d1QW146",
        "outputId": "c56bd5ab-9aa1-45b2-db62-1012096becc0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+----------+------------------+--------------------+\n",
            "|life_point|          variable|               value|\n",
            "+----------+------------------+--------------------+\n",
            "|        12|      avg_activity|  0.2387706855791962|\n",
            "|        12|          avg_food|   0.284160756501182|\n",
            "|        12|        avg_people| 0.04846335697399527|\n",
            "|        12|         avg_event|0.031914893617021274|\n",
            "|        12|        avg_travel|0.040780141843971635|\n",
            "|        12|avg_transportation| 0.08687943262411348|\n",
            "|        12|       avg_utility|  0.1756501182033097|\n",
            "|        12|          avg_cash|  0.0425531914893617|\n",
            "|        12|       avg_illegal|0.050827423167848704|\n",
            "|         1|      avg_activity| 0.09090909090909091|\n",
            "|         1|          avg_food|                 0.5|\n",
            "|         1|        avg_people|                 0.0|\n",
            "|         1|         avg_event|                 0.0|\n",
            "|         1|        avg_travel|0.045454545454545456|\n",
            "|         1|avg_transportation|  0.0303030303030303|\n",
            "|         1|       avg_utility|  0.1515151515151515|\n",
            "|         1|          avg_cash| 0.09090909090909091|\n",
            "|         1|       avg_illegal| 0.09090909090909091|\n",
            "|         6|      avg_activity| 0.17307692307692307|\n",
            "|         6|          avg_food| 0.32905982905982906|\n",
            "+----------+------------------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import expr, stddev\n",
        "\n",
        "std_by_time_df = classification_by_time.groupBy(\"life_point\").agg(\n",
        "    stddev(\"count_activity\").alias(\"std_activity\"),\n",
        "    stddev(\"count_food\").alias(\"std_food\"),\n",
        "    stddev(\"count_people\").alias(\"std_people\"),\n",
        "    stddev(\"count_event\").alias(\"std_event\"),\n",
        "    stddev(\"count_travel\").alias(\"std_travel\"),\n",
        "    stddev(\"count_transportation\").alias(\"std_transportation\"),\n",
        "    stddev(\"count_utility\").alias(\"std_utility\"),\n",
        "    stddev(\"count_cash\").alias(\"std_cash\"),\n",
        "    stddev(\"count_illegal\").alias(\"std_illegal\")\n",
        ")\n",
        "\n",
        "melted_columns = ['std_activity', 'std_food', 'std_people', 'std_event', 'std_travel', 'std_transportation',\n",
        "                  'std_utility', 'std_cash', 'std_illegal']\n",
        "\n",
        "melted_columns_expr = \", \".join([\"'{0}', {0}\".format(col) for col in melted_columns])\n",
        "\n",
        "std_by_time_df = std_by_time_df.selectExpr(\"life_point\", \"stack({0}, {1}) as (variable, value)\".format(\n",
        "    len(melted_columns), melted_columns_expr\n",
        "))\n",
        "\n",
        "std_by_time_df.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wUK-CoL8bgMc",
        "outputId": "cdc88096-9b3c-47a4-b642-7df1b4871647"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+----------+------------------+-------------------+\n",
            "|life_point|          variable|              value|\n",
            "+----------+------------------+-------------------+\n",
            "|        12|      std_activity| 0.4080319415110253|\n",
            "|        12|          std_food| 0.4350439051056599|\n",
            "|        12|        std_people|0.19874507789871862|\n",
            "|        12|         std_event|0.14895787944256025|\n",
            "|        12|        std_travel|0.19045085951119037|\n",
            "|        12|std_transportation|0.28028336363598955|\n",
            "|        12|       std_utility|  0.370368315878575|\n",
            "|        12|          std_cash|0.18409393663996323|\n",
            "|        12|       std_illegal|0.20887383449025324|\n",
            "|         1|      std_activity|0.30151134457776363|\n",
            "|         1|          std_food|                0.5|\n",
            "|         1|        std_people|                0.0|\n",
            "|         1|         std_event|                0.0|\n",
            "|         1|        std_travel|0.15075567228888181|\n",
            "|         1|std_transportation|0.10050378152592121|\n",
            "|         1|       std_utility|0.34524182711820883|\n",
            "|         1|          std_cash|0.30151134457776363|\n",
            "|         1|       std_illegal|0.30151134457776363|\n",
            "|         6|      std_activity|  0.349594907271664|\n",
            "|         6|          std_food| 0.4627769906876437|\n",
            "+----------+------------------+-------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "avg_by_time_pd = avg_by_time_df.toPandas()\n",
        "std_by_time_pd = std_by_time_df.toPandas()"
      ],
      "metadata": {
        "id": "XV09wSkdcdGi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "\n",
        "# Define the variables and subplots\n",
        "variables = ['activity', 'people', 'travel', 'transportation', 'food', 'illegal', 'cash', 'utility', 'event']\n",
        "fig, axes = plt.subplots(3, 3, figsize=(15, 10))\n",
        "plt.subplots_adjust(hspace=0.4)\n",
        "sns.despine(left=True)\n",
        "\n",
        "# Iterate over variables and create subplots\n",
        "for idx, var in enumerate(variables):\n",
        "    avg_variable = 'avg_' + var\n",
        "    std_variable = 'std_' + var\n",
        "\n",
        "    x = []\n",
        "    lower_bound = []\n",
        "    upper_bound = []\n",
        "\n",
        "    for point in range(1, 13):\n",
        "        avg = avg_by_time_pd.loc[(avg_by_time_pd.variable == avg_variable) & (avg_by_time_pd.life_point == point), 'value'].values[0]\n",
        "        std = std_by_time_pd.loc[(std_by_time_pd.variable == std_variable) & (std_by_time_pd.life_point == point), 'value'].values[0]\n",
        "\n",
        "        x.append(avg)\n",
        "        lower_bound.append(avg - 2 * std)\n",
        "        upper_bound.append(avg + 2 * std)\n",
        "\n",
        "    y = list(range(1, 13))\n",
        "\n",
        "    # Plot the confidence interval\n",
        "    axes[idx // 3, idx % 3].set_xlim(-0.75, 1)\n",
        "    axes[idx // 3, idx % 3].fill_betweenx(y, lower_bound, upper_bound, alpha=0.3)\n",
        "\n",
        "    # Plot the average\n",
        "    axes[idx // 3, idx % 3].scatter(x, y)\n",
        "\n",
        "    axes[idx // 3, idx % 3].set_title(avg_variable)\n",
        "    axes[idx // 3, idx % 3].set_ylabel(\"lifetime in months\")\n",
        "\n",
        "# Show the plots\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 853
        },
        "id": "BvHEC_oGcnAC",
        "outputId": "249a7192-7539-4215-c233-effe74ee867b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x1000 with 9 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Most of the average spendings stabilize after the first life point."
      ],
      "metadata": {
        "id": "w5GpefFTdFO-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Social Network Analytics (SNA)"
      ],
      "metadata": {
        "id": "512LceTuQKcv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark = SparkSession\\\n",
        "        .builder\\\n",
        "        .master(\"local[*]\")\\\n",
        "        .config(\"spark.jars.packages\", \"graphframes:graphframes:0.8.1-spark3.0-s_2.12\")\\\n",
        "        .getOrCreate()"
      ],
      "metadata": {
        "id": "r31wGKorQ4lF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "venmo_df = venmo_df.sample(fraction=0.01, seed=42)"
      ],
      "metadata": {
        "id": "zhfe2SbGRFBw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a3espp4rxSRy",
        "outputId": "d345f256-fe55-4a82-b149-5301938ec9db"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+--------+----------------+--------------------+--------------------+-----------+--------------------+\n",
            "|  user1|   user2|transaction_type|            datetime|         description|is_business|            story_id|\n",
            "+-------+--------+----------------+--------------------+--------------------+-----------+--------------------+\n",
            "| 473246|  357995|         payment|2016-02-10T08:58:...|                🎅🏿|      false|56ba8b29cd03c9af2...|\n",
            "| 811090|  864824|          charge|2014-12-12T15:17:...|          Zp dinnah!|      false|548a9684ca134c404...|\n",
            "|1508991| 1141580|         payment|2015-01-27T01:46:...|            🍻🍻🍺🍺|      false|54c67d5bca134c404...|\n",
            "|3959333| 2942063|         payment|2016-02-08T23:32:...|              🏡🍾👭|      false|56b8b510cd03c9af2...|\n",
            "|1069346| 3543226|          charge|2015-10-07T10:26:...|        Pay yo bills|      false|561490e5cd03c9af2...|\n",
            "|5621067| 7237863|         payment|2016-05-21T04:54:...|                  📺|      false|573f8799cd03c9af2...|\n",
            "|2164273|  587489|         payment|2015-07-05T11:29:...|              Stuffs|      false|5598b2a607f81c216...|\n",
            "| 550288|  327652|         payment|2013-09-30T23:08:...|       Reimbursement|      false|5249a1f0d56b6bac5...|\n",
            "|7416216|11587064|         payment|2016-07-22T07:32:...|        Thanks again|      false|579169a923e064eac...|\n",
            "| 737239|   77724|          charge|2014-08-31T13:29:...|               Magic|      false|5402c0e77d0b0354e...|\n",
            "|2728526| 3077414|         payment|2015-08-20T08:01:...|                👧🏿|      false|55d526dc5d6cc8730...|\n",
            "|1501871| 1099703|         payment|2015-04-02T06:14:...|           Volcanoes|      false|551c7be607f81c4b9...|\n",
            "|3410728| 2233850|         payment|2016-01-31T14:09:...|                  🙃|      false|56ada514cd03c9af2...|\n",
            "|1474097| 1874860|         payment|2015-03-11T23:22:...|       Frank n beans|      false|55006bd2cd03c9af2...|\n",
            "| 677727| 7611964|         payment|2015-12-03T10:05:...|         Candy grams|      false|565fa37dcd03c9af2...|\n",
            "|1224621| 1223874|         payment|2015-10-07T03:21:...|              ✈️🎃😈|      false|56142d2ecd03c9af2...|\n",
            "|1040339|  338990|         payment|2015-11-14T11:48:...|             Last ga|      false|5646af1ccd03c9af2...|\n",
            "|2429984| 1094020|         payment|2014-10-02T22:37:...|   for pickem week 5|      false|542d71367d0b0354e...|\n",
            "|4576935| 8624986|         payment|2016-07-08T22:39:...|                  🏫|      false|577fc91823e064eac...|\n",
            "|2580631| 8280398|          charge|2016-04-01T00:48:...|Brand new HP Lapt...|      false|56fd6309cd03c9af2...|\n",
            "+-------+--------+----------------+--------------------+--------------------+-----------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ],
      "source": [
        "venmo_df.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DuDoHGujPoPp"
      },
      "source": [
        "### Script to find user’s friends and friends of friends\n",
        "\n",
        "(Friend definition: A user’s friend is someone who has transacted with the user, either sending money to the user or receiving money from the user).\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KdymH-t4HAmE",
        "outputId": "0d8dac60-4190-471b-8201-134d7004739a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+\n",
            "|  user1|  user2|\n",
            "+-------+-------+\n",
            "|1542962|1561864|\n",
            "|2625628|2626024|\n",
            "|2918884|2244083|\n",
            "| 696661| 198957|\n",
            "| 304267| 276828|\n",
            "|2737539|3002157|\n",
            "|2991028|4114608|\n",
            "|8977153|6819793|\n",
            "|1541941|4726257|\n",
            "|1879914|2950156|\n",
            "|1452517| 525298|\n",
            "| 139135|7819027|\n",
            "|2657367|6903077|\n",
            "|5608012|4441477|\n",
            "|2717207|1552945|\n",
            "|6203804|6252525|\n",
            "| 965058| 221317|\n",
            "|5766322|1789988|\n",
            "|5374147|7286662|\n",
            "|2140410|2450827|\n",
            "+-------+-------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ],
      "source": [
        "user_subset = venmo_df.select(\"user1\", \"user2\").dropDuplicates()\n",
        "# Create a temporary view for the user_subset\n",
        "user_subset.createOrReplaceTempView(\"user_subset\")\n",
        "user_subset.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### List of friends and friends of friends"
      ],
      "metadata": {
        "id": "Xj9AFDylu30Y"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Find friends for all users\n",
        "query = \"\"\"\n",
        "          SELECT A.user1 AS user_id, A.user2 AS friend_id\n",
        "          FROM user_subset A\n",
        "          UNION\n",
        "          SELECT B.user2 AS user_id, B.user1 AS friend_id\n",
        "          FROM user_subset B\n",
        "\"\"\"\n",
        "\n",
        "user_friends = spark.sql(query)\n",
        "user_friends.createOrReplaceTempView(\"user_friends\")\n",
        "user_friends.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pzC3ZbjJQrvC",
        "outputId": "14806d33-cff4-4eae-abce-fdd2216200ea"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+\n",
            "|user_id|friend_id|\n",
            "+-------+---------+\n",
            "|1542962|  1561864|\n",
            "|2625628|  2626024|\n",
            "|2918884|  2244083|\n",
            "| 696661|   198957|\n",
            "| 304267|   276828|\n",
            "|2737539|  3002157|\n",
            "|2991028|  4114608|\n",
            "|8977153|  6819793|\n",
            "|1541941|  4726257|\n",
            "|1879914|  2950156|\n",
            "|1452517|   525298|\n",
            "| 139135|  7819027|\n",
            "|2657367|  6903077|\n",
            "|5608012|  4441477|\n",
            "|2717207|  1552945|\n",
            "|6203804|  6252525|\n",
            "| 965058|   221317|\n",
            "|5766322|  1789988|\n",
            "|5374147|  7286662|\n",
            "|2140410|  2450827|\n",
            "+-------+---------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_RqRuhhL-z_v",
        "outputId": "9be9db29-213a-4716-c6e5-3c67d1d39490"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+-----------+--------------+\n",
            "| user_id|    friends|friends_friend|\n",
            "+--------+-----------+--------------+\n",
            "|      10|[71056, 43]|       [47104]|\n",
            "| 1000013|   [882320]|          null|\n",
            "| 1000024|  [1571549]|      [691294]|\n",
            "| 1000031|  [3692874]|          null|\n",
            "| 1000063|   [999123]|          null|\n",
            "|10000740|  [6286189]|          null|\n",
            "| 1000099|  [3149184]|          null|\n",
            "| 1000119|  [1562407]|          null|\n",
            "| 1000129|  [2939678]|          null|\n",
            "| 1000132|  [4300675]|          null|\n",
            "| 1000158|   [511362]|          null|\n",
            "|10001662|  [3759599]|          null|\n",
            "| 1000204|  [4215633]|          null|\n",
            "| 1000222|   [886822]|          null|\n",
            "|10002713|  [8035734]|          null|\n",
            "| 1000278|  [1716420]|     [1918994]|\n",
            "| 1000281|  [5774380]|          null|\n",
            "|10003088|  [1413336]|          null|\n",
            "| 1000318|  [1525495]|      [760137]|\n",
            "|10003386|  [3552552]|          null|\n",
            "+--------+-----------+--------------+\n",
            "only showing top 20 rows\n",
            "\n",
            "time: 15.5 s (started: 2023-05-25 00:26:36 +00:00)\n"
          ]
        }
      ],
      "source": [
        "# Find friends for all users\n",
        "query = \"\"\"\n",
        "          with temp1 as(\n",
        "          SELECT A.user1 AS user_id, A.user2 AS friend_id, 1 AS Contact_Num\n",
        "          FROM user_subset A\n",
        "          UNION\n",
        "          SELECT B.user2 AS user_id, B.user1 AS friend_id, 1 As Contact_Num\n",
        "          FROM user_subset B\n",
        "          ),\n",
        "          temp2 as (\n",
        "          Select distinct A.user_id as user_id,\n",
        "          B.friend_id as friend_id, 2 AS Contact_Num\n",
        "          From temp1 A\n",
        "          Left JOIN temp1 B on A.friend_id = B.user_id\n",
        "          and A.user_id <> B.friend_id\n",
        "          where B.friend_id is not NULL\n",
        "          ),\n",
        "          temp3 as (\n",
        "          Select user_id, friend_id, Contact_Num\n",
        "          from temp1\n",
        "          UNION\n",
        "          Select user_id, friend_id, Contact_Num\n",
        "          from temp2\n",
        "          order by user_id\n",
        "          ),\n",
        "          temp4 as (\n",
        "          Select *,\n",
        "          row_number() OVER(PARTITION BY user_id, friend_id ORDER BY Contact_Num) AS rank_num\n",
        "          FROM temp3\n",
        "          ),\n",
        "          cte_f as (\n",
        "          select user_id,\n",
        "            collect_list(friend_id) as friends\n",
        "            from temp1\n",
        "            group by user_id\n",
        "            ),\n",
        "          cte_ff (\n",
        "          Select user_id, collect_list(friend_id) as friends_friend\n",
        "          FROM temp4\n",
        "          where rank_num =1 and Contact_Num = 2\n",
        "          group by user_id\n",
        "          )\n",
        "          select cte_f.user_id, friends, friends_friend\n",
        "          From cte_f\n",
        "          Left join cte_ff on cte_f.user_id = cte_ff.user_id\n",
        "          order by  user_id\n",
        "\"\"\"\n",
        "\n",
        "all_friends = spark.sql(query)\n",
        "all_friends.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s_B8_IwdPHzy"
      },
      "source": [
        "## Social network metrics across a user’s lifetime in Venmo (from 0 up to 12 months)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Number of friends and number of friends of friends"
      ],
      "metadata": {
        "id": "RBzFWMgDTVPZ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Static count of Friends and Friends of Friends"
      ],
      "metadata": {
        "id": "kAuUmj5JYFdJ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CZoHEIvxJTal",
        "outputId": "7e495cd9-aaf3-4e6a-df3f-775fce365aac"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "+--------+-----------+--------------+-----+\n",
            "| user_id|Contact_Num|       Contact|Total|\n",
            "+--------+-----------+--------------+-----+\n",
            "|      10|          1|        Friend|    3|\n",
            "|      10|          2|FriendOfFriend|    2|\n",
            "| 1000002|          1|        Friend|    1|\n",
            "|10000027|          1|        Friend|    1|\n",
            "| 1000006|          1|        Friend|    1|\n",
            "| 1000006|          2|FriendOfFriend|    1|\n",
            "| 1000013|          1|        Friend|    4|\n",
            "| 1000013|          2|FriendOfFriend|    5|\n",
            "|10000139|          1|        Friend|    1|\n",
            "|10000139|          2|FriendOfFriend|    1|\n",
            "| 1000022|          1|        Friend|    3|\n",
            "| 1000022|          2|FriendOfFriend|    7|\n",
            "| 1000024|          1|        Friend|    2|\n",
            "| 1000024|          2|FriendOfFriend|    2|\n",
            "|10000289|          1|        Friend|    1|\n",
            "| 1000029|          1|        Friend|    1|\n",
            "| 1000031|          1|        Friend|    3|\n",
            "| 1000031|          2|FriendOfFriend|    6|\n",
            "| 1000035|          1|        Friend|    2|\n",
            "| 1000035|          2|FriendOfFriend|    1|\n",
            "+--------+-----------+--------------+-----+\n",
            "only showing top 20 rows\n",
            "\n",
            "time: 2min 10s (started: 2023-05-17 00:22:29 +00:00)\n"
          ]
        }
      ],
      "source": [
        "# Find friends for all users static\n",
        "query = \"\"\"\n",
        "          with temp1 as(\n",
        "          SELECT A.user1 AS user_id, A.user2 AS friend_id, 1 AS Contact_Num\n",
        "          FROM user_subset A\n",
        "          UNION\n",
        "          SELECT B.user2 AS user_id, B.user1 AS friend_id, 1 As Contact_Num\n",
        "          FROM user_subset B\n",
        "          ),\n",
        "          temp2 as (\n",
        "          Select distinct A.user_id as user_id,\n",
        "          B.friend_id as friend_id, 2 AS Contact_Num\n",
        "          From temp1 A\n",
        "          Left JOIN temp1 B on A.friend_id = B.user_id\n",
        "          and A.user_id <> B.friend_id\n",
        "          where B.friend_id is not NULL\n",
        "          ),\n",
        "          temp3 as (\n",
        "          Select user_id, friend_id, Contact_Num\n",
        "          from temp1\n",
        "          UNION\n",
        "          Select user_id, friend_id, Contact_Num\n",
        "          from temp2\n",
        "          order by user_id\n",
        "          ),\n",
        "          temp4 as (\n",
        "          Select *,\n",
        "          row_number() OVER(PARTITION BY user_id, friend_id ORDER BY Contact_Num) AS rank_num\n",
        "          FROM temp3\n",
        "          )\n",
        "          Select user_id,\n",
        "          CASE\n",
        "          WHEN Contact_Num = 1 THEN 'Friend'\n",
        "          WHEN Contact_Num = 2 THEN 'FriendOfFriend'\n",
        "          END AS Contact,\n",
        "          sum(rank_num) as Total\n",
        "          from temp4\n",
        "          where rank_num =1\n",
        "          group by user_id, Contact_Num\n",
        "          order by user_id, Contact_Num\n",
        "\"\"\"\n",
        "\n",
        "static_friends_count = spark.sql(query)\n",
        "static_friends_count.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Lifetime count of Friends and Friends of Friends"
      ],
      "metadata": {
        "id": "iqfDuqXmYNF5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "user_month = venmo_df.select(\"user1\", \"user2\",\"datetime\").dropDuplicates()\n",
        "# Create a temporary view for the user_subset\n",
        "user_month.createOrReplaceTempView(\"user_month\")\n",
        "user_month.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_ebuhuKGvbJn",
        "outputId": "543b826b-d056-4189-a4b3-32911e0c4061"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+-------+--------------------+\n",
            "|  user1|  user2|            datetime|\n",
            "+-------+-------+--------------------+\n",
            "|3245723|1951007|2016-08-07T03:50:...|\n",
            "|1688153|1535384|2015-04-26T00:46:...|\n",
            "|1255452|1136797|2014-04-22T07:42:...|\n",
            "|2833021|1448124|2016-10-16T11:54:...|\n",
            "|2968160|1867209|2016-07-10T09:19:...|\n",
            "|8128981|5315878|2016-03-04T05:28:...|\n",
            "| 830284| 830555|2015-03-08T10:16:...|\n",
            "|3026689| 492872|2015-04-17T00:53:...|\n",
            "|1772857|1476983|2015-08-12T01:52:...|\n",
            "|2208210|1915443|2015-10-16T06:30:...|\n",
            "|1849616|1853994|2014-10-06T05:01:...|\n",
            "|7100227|2355572|2016-03-10T08:14:...|\n",
            "|2897725| 922630|2015-07-24T06:25:...|\n",
            "|1124952| 889605|2016-08-09T11:15:...|\n",
            "|2237730|2293597|2016-05-08T23:40:...|\n",
            "|2470605| 661130|2016-01-28T03:20:...|\n",
            "|4152388|6631921|2015-11-18T16:04:...|\n",
            "| 527265| 640428|2015-05-14T23:01:...|\n",
            "| 959416|1301570|2015-10-20T00:32:...|\n",
            "|6138527|9464194|2016-10-01T05:42:...|\n",
            "+-------+-------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Count friends for all users monthwise\n",
        "query = \"\"\" with temp1 as (\n",
        "          Select distinct user1 as user_id, user2 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "          Union\n",
        "          Select distinct user2 as user_id, user1 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "        ),\n",
        "        temp2 as (\n",
        "          Select user_id, min(datetime) as firstDate\n",
        "          From temp1\n",
        "          group by user_id\n",
        "        ),\n",
        "        temp3 as (\n",
        "          Select temp1.user_id, friend_id, datetime, firstDate,\n",
        "          DATEDIFF(month, firstDate, datetime) as month_num,\n",
        "          dense_rank() OVER(PARTITION BY temp1.user_id, friend_id ORDER BY datetime) AS rank_num\n",
        "          From temp1\n",
        "          Left Join temp2 on temp1.user_id = temp2.user_id\n",
        "          where DATEDIFF(month, firstDate, datetime) between 0 and 12\n",
        "        )\n",
        "          Select distinct user_id, month_num,\n",
        "          count(friend_id) OVER(PARTITION BY user_id, datetime) AS friends_count\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "          order by user_id, month_num\n",
        "\"\"\"\n",
        "\n",
        "dynamic_friends_count = spark.sql(query)\n",
        "dynamic_friends_count.createOrReplaceTempView(\"dynamic_friends_count\")\n",
        "dynamic_friends_count.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v_ju6AjTQfis",
        "outputId": "7a28e165-0f03-448c-ce37-15122e7ec8e5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+---------+-------------+\n",
            "| user_id|month_num|friends_count|\n",
            "+--------+---------+-------------+\n",
            "|      10|        0|            1|\n",
            "|      10|       12|            1|\n",
            "| 1000013|        0|            1|\n",
            "| 1000024|        0|            1|\n",
            "| 1000031|        0|            1|\n",
            "| 1000063|        0|            1|\n",
            "|10000740|        0|            1|\n",
            "| 1000099|        0|            1|\n",
            "| 1000119|        0|            1|\n",
            "| 1000129|        0|            1|\n",
            "| 1000132|        0|            1|\n",
            "| 1000158|        0|            1|\n",
            "|10001662|        0|            1|\n",
            "| 1000204|        0|            1|\n",
            "| 1000222|        0|            1|\n",
            "|10002713|        0|            1|\n",
            "| 1000278|        0|            1|\n",
            "| 1000281|        0|            1|\n",
            "|10003088|        0|            1|\n",
            "| 1000318|        0|            1|\n",
            "+--------+---------+-------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Count friends of friends for all users\n",
        "query = \"\"\" with temp1 as (\n",
        "          Select distinct user1 as user_id, user2 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "          Union\n",
        "          Select distinct user2 as user_id, user1 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "        ),\n",
        "        temp2 as (\n",
        "          Select user_id, min(datetime) as firstDate\n",
        "          From temp1\n",
        "          group by user_id\n",
        "        ),\n",
        "        temp3 as (\n",
        "          Select temp1.user_id as user_id, friend_id, datetime, firstDate,\n",
        "          DATEDIFF(month, firstDate, datetime) as month_num,\n",
        "          dense_rank() OVER(PARTITION BY temp1.user_id, friend_id ORDER BY datetime) AS rank_num\n",
        "          From temp1\n",
        "          Left Join temp2 on temp1.user_id = temp2.user_id\n",
        "          where DATEDIFF(month, firstDate, datetime) between 0 and 12\n",
        "        ),\n",
        "        temp4 as (\n",
        "          Select distinct user_id, friend_id, month_num, 1 As Contact_Num\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "          UNION\n",
        "          Select distinct A.user_id as user_id,\n",
        "          B.friend_id as friend_id,\n",
        "          DATEDIFF(month, B.datetime, A.firstDate) as month_num,\n",
        "          2 AS Contact_Num\n",
        "          From temp3 A\n",
        "          Left JOIN temp3 B on A.friend_id = B.user_id\n",
        "          and A.user_id <> B.friend_id\n",
        "          where B.friend_id is not NULL\n",
        "          and DATEDIFF(month, B.datetime, A.firstDate) between 0 and 12\n",
        "        ),\n",
        "        temp5 as (\n",
        "          Select *,\n",
        "          row_number() OVER(PARTITION BY user_id, friend_id, month_num ORDER BY Contact_Num) AS rank_num\n",
        "          FROM temp4\n",
        "        )\n",
        "          Select user_id, month_num,\n",
        "          count(friend_id) OVER(PARTITION BY user_id, Contact_Num, month_num) AS friends_friends_count\n",
        "          from temp5\n",
        "          where rank_num=1 and Contact_Num = 2\n",
        "          order by user_id, month_num\n",
        "\n",
        "\"\"\"\n",
        "\n",
        "dynamic_fof_count = spark.sql(query)\n",
        "dynamic_fof_count.createOrReplaceTempView(\"dynamic_fof_count\")\n",
        "dynamic_fof_count.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qA5z9nmmeGR1",
        "outputId": "66669104-8696-4011-e05d-bf4d3b330460"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+---------+---------------------+\n",
            "| user_id|month_num|friends_friends_count|\n",
            "+--------+---------+---------------------+\n",
            "|      10|        6|                    1|\n",
            "| 1000024|        8|                    1|\n",
            "| 1000278|        6|                    1|\n",
            "| 1000318|        2|                    1|\n",
            "| 1001206|        1|                    1|\n",
            "| 1001264|        4|                    1|\n",
            "|10027969|        3|                    1|\n",
            "| 1002827|        6|                    1|\n",
            "| 1003416|        0|                    1|\n",
            "| 1003671|        7|                    1|\n",
            "| 1003971|       12|                    1|\n",
            "|10040360|       10|                    1|\n",
            "|  100490|       10|                    1|\n",
            "| 1006114|        6|                    1|\n",
            "|10061224|        7|                    1|\n",
            "|10066972|        0|                    1|\n",
            "| 1006833|        6|                    1|\n",
            "| 1007396|        6|                    1|\n",
            "| 1008961|        2|                    1|\n",
            "|    1009|        0|                    1|\n",
            "+--------+---------+---------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# create continuous user lifetime\n",
        "month_list = [0,1,2,3,4,5,6,7,8,9,10,11,12]\n",
        "dummy_month = user_friends.select(\"user_id\").distinct().withColumn(\"month_num\", F.array([F.lit(x) for x in month_list]))\n",
        "dummy_month = dummy_month.withColumn(\"month_num\", F.explode(col(\"month_num\"))).orderBy(\"user_id\",\"month_num\")\n",
        "dummy_month.createOrReplaceTempView(\"dummy_month\")\n",
        "dummy_month.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b6VsPZ6CQdTm",
        "outputId": "e5726910-18f9-4950-e9bb-1e43d7dc95ea"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+\n",
            "|user_id|month_num|\n",
            "+-------+---------+\n",
            "|     10|        0|\n",
            "|     10|        1|\n",
            "|     10|        2|\n",
            "|     10|        3|\n",
            "|     10|        4|\n",
            "|     10|        5|\n",
            "|     10|        6|\n",
            "|     10|        7|\n",
            "|     10|        8|\n",
            "|     10|        9|\n",
            "|     10|       10|\n",
            "|     10|       11|\n",
            "|     10|       12|\n",
            "|1000013|        0|\n",
            "|1000013|        1|\n",
            "|1000013|        2|\n",
            "|1000013|        3|\n",
            "|1000013|        4|\n",
            "|1000013|        5|\n",
            "|1000013|        6|\n",
            "+-------+---------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Count friends of friends for all users\n",
        "query = \"\"\"\n",
        "        with temp1 as (\n",
        "        Select dummy_month.user_id, dummy_month.month_num\n",
        "        ,COALESCE(dynamic_friends_count.friends_count, 0) AS friends_count\n",
        "        ,COALESCE(dynamic_fof_count.friends_friends_count, 0) AS friends_friends_count\n",
        "        From dummy_month\n",
        "        Left JOIN dynamic_friends_count on dynamic_friends_count.user_id = dummy_month.user_id and dynamic_friends_count.month_num = dummy_month.month_num\n",
        "        Left JOIN dynamic_fof_count ON dynamic_fof_count.user_id = dummy_month.user_id and dynamic_fof_count.month_num = dummy_month.month_num\n",
        "        order by user_id, month_num\n",
        "        )\n",
        "        Select user_id, month_num\n",
        "        ,sum(friends_count) OVER(PARTITION BY user_id order by month_num) AS friends_count_cum\n",
        "        ,sum(friends_friends_count) OVER(PARTITION BY user_id order by month_num) AS friends_friends_count_cum\n",
        "        from temp1\n",
        "        order by user_id, month_num\n",
        "\"\"\"\n",
        "\n",
        "df_f_fof = spark.sql(query)\n",
        "df_f_fof.createOrReplaceTempView(\"df_f_fof\")\n",
        "df_f_fof.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aLs3xuKxPy-G",
        "outputId": "84d848f7-a96e-4c57-fdc6-cde52a9ce415"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|\n",
            "+-------+---------+-----------------+-------------------------+\n",
            "|     10|        0|                1|                        0|\n",
            "|     10|        1|                1|                        0|\n",
            "|     10|        2|                1|                        0|\n",
            "|     10|        3|                1|                        0|\n",
            "|     10|        4|                1|                        0|\n",
            "|     10|        5|                1|                        0|\n",
            "|     10|        6|                1|                        1|\n",
            "|     10|        7|                1|                        1|\n",
            "|     10|        8|                1|                        1|\n",
            "|     10|        9|                1|                        1|\n",
            "|     10|       10|                1|                        1|\n",
            "|     10|       11|                1|                        1|\n",
            "|     10|       12|                2|                        1|\n",
            "|1000013|        0|                1|                        0|\n",
            "|1000013|        1|                1|                        0|\n",
            "|1000013|        2|                1|                        0|\n",
            "|1000013|        3|                1|                        0|\n",
            "|1000013|        4|                1|                        0|\n",
            "|1000013|        5|                1|                        0|\n",
            "|1000013|        6|                1|                        0|\n",
            "+-------+---------+-----------------+-------------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Lifetime List of Friends and Friends of Friends"
      ],
      "metadata": {
        "id": "Jnd6WPpIvy2z"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# List friends for all users monthwise\n",
        "query = \"\"\" with temp1 as (\n",
        "          Select distinct user1 as user_id, user2 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "          Union\n",
        "          Select distinct user2 as user_id, user1 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "        ),\n",
        "        temp2 as (\n",
        "          Select user_id, min(datetime) as firstDate\n",
        "          From temp1\n",
        "          group by user_id\n",
        "        ),\n",
        "        temp3 as (\n",
        "          Select temp1.user_id, friend_id, datetime, firstDate,\n",
        "          DATEDIFF(month, firstDate, datetime) as month_num,\n",
        "          dense_rank() OVER(PARTITION BY temp1.user_id, friend_id ORDER BY datetime) AS rank_num\n",
        "          From temp1\n",
        "          Left Join temp2 on temp1.user_id = temp2.user_id\n",
        "          where DATEDIFF(month, firstDate, datetime) <=12\n",
        "        ),\n",
        "        temp4 as (\n",
        "          Select distinct user_id,\n",
        "          month_num,\n",
        "          friend_id\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "        )\n",
        "        Select user_id, month_num,\n",
        "        collect_list(friend_id) OVER(PARTITION BY user_id ORDER BY month_num) as friend_id_list\n",
        "        from temp4\n",
        "        order by user_id, month_num\n",
        "\n",
        "\n",
        "\"\"\"\n",
        "\n",
        "dynamic_friends_list = spark.sql(query)\n",
        "dynamic_friends_list.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bRaXhgyJpaUG",
        "outputId": "33f9b514-b3e1-429b-9639-2a4f1233a8c9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+---------+--------------+\n",
            "| user_id|month_num|friend_id_list|\n",
            "+--------+---------+--------------+\n",
            "|      10|        0|       [71056]|\n",
            "|      10|       12|   [71056, 43]|\n",
            "| 1000013|        0|      [882320]|\n",
            "| 1000024|        0|     [1571549]|\n",
            "| 1000031|        0|     [3692874]|\n",
            "| 1000063|        0|      [999123]|\n",
            "|10000740|        0|     [6286189]|\n",
            "| 1000099|        0|     [3149184]|\n",
            "| 1000119|        0|     [1562407]|\n",
            "| 1000129|        0|     [2939678]|\n",
            "| 1000132|        0|     [4300675]|\n",
            "| 1000158|        0|      [511362]|\n",
            "|10001662|        0|     [3759599]|\n",
            "| 1000204|        0|     [4215633]|\n",
            "| 1000222|        0|      [886822]|\n",
            "|10002713|        0|     [8035734]|\n",
            "| 1000278|        0|     [1716420]|\n",
            "| 1000281|        0|     [5774380]|\n",
            "|10003088|        0|     [1413336]|\n",
            "| 1000318|        0|     [1525495]|\n",
            "+--------+---------+--------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# List friends of friends for all users monthwise\n",
        "query = \"\"\" with temp1 as (\n",
        "          Select distinct user1 as user_id, user2 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "          Union\n",
        "          Select distinct user2 as user_id, user1 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "        ),\n",
        "        temp2 as (\n",
        "          Select user_id, min(datetime) as firstDate\n",
        "          From temp1\n",
        "          group by user_id\n",
        "        ),\n",
        "        temp3 as (\n",
        "          Select temp1.user_id as user_id, friend_id, datetime, firstDate,\n",
        "          DATEDIFF(month, firstDate, datetime) as month_num,\n",
        "          dense_rank() OVER(PARTITION BY temp1.user_id, friend_id ORDER BY datetime) AS rank_num\n",
        "          From temp1\n",
        "          Left Join temp2 on temp1.user_id = temp2.user_id\n",
        "          where DATEDIFF(month, firstDate, datetime) between 0 and 12\n",
        "        ),\n",
        "        temp4 as (\n",
        "          Select distinct user_id, friend_id, month_num, 1 As Contact_Num\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "          UNION\n",
        "          Select distinct A.user_id as user_id,\n",
        "          B.friend_id as friend_id,\n",
        "          DATEDIFF(month, B.datetime, A.firstDate) as month_num,\n",
        "          2 AS Contact_Num\n",
        "          From temp3 A\n",
        "          Left JOIN temp3 B on A.friend_id = B.user_id\n",
        "          and A.user_id <> B.friend_id\n",
        "          where B.friend_id is not NULL\n",
        "          and DATEDIFF(month, B.datetime, A.firstDate) between 0 and 12\n",
        "        ),\n",
        "        temp5 as (\n",
        "          Select *,\n",
        "          row_number() OVER(PARTITION BY user_id, friend_id, month_num ORDER BY Contact_Num) AS rank_num\n",
        "          FROM temp4\n",
        "        ),\n",
        "        temp6 as (\n",
        "          Select user_id, month_num, friend_id\n",
        "          from temp5\n",
        "          where rank_num=1 and Contact_Num = 2\n",
        "        )\n",
        "        Select user_id, month_num,\n",
        "        collect_list(friend_id) OVER(PARTITION BY user_id order by month_num) AS friends_friends_list\n",
        "        from temp6\n",
        "        order by user_id, month_num\n",
        "\n",
        "\"\"\"\n",
        "\n",
        "dynamic_fof_list = spark.sql(query)\n",
        "dynamic_fof_list.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EfW6c6lIwCSv",
        "outputId": "9829a4e2-af8c-451f-e2c0-2ce527d2ac9a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+---------+--------------------+\n",
            "| user_id|month_num|friends_friends_list|\n",
            "+--------+---------+--------------------+\n",
            "| 1007396|        4|           [1586113]|\n",
            "|10279155|        3|           [2281739]|\n",
            "|10809899|        9|           [2285029]|\n",
            "| 1083330|        9|            [130950]|\n",
            "| 1084455|        5|            [397100]|\n",
            "| 1109017|        2|            [953789]|\n",
            "| 1111929|        3|           [1474414]|\n",
            "|11168716|       11|           [6248339]|\n",
            "| 1117183|        9|           [1995845]|\n",
            "| 1172062|        3|            [964699]|\n",
            "| 1179249|        1|            [535118]|\n",
            "| 1202240|        6|           [1744728]|\n",
            "| 1229847|        0|            [669367]|\n",
            "| 1232856|       12|           [2338080]|\n",
            "| 1243976|       10|           [3551154]|\n",
            "| 1246917|       10|           [2611312]|\n",
            "| 1344168|        7|           [1347032]|\n",
            "| 1375825|        6|           [3189875]|\n",
            "| 1378334|        5|           [5583707]|\n",
            "| 1383318|       12|            [987554]|\n",
            "+--------+---------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# List friends of friends for all users monthwise\n",
        "query = \"\"\" with temp1 as (\n",
        "          Select distinct user1 as user_id, user2 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "          Union\n",
        "          Select distinct user2 as user_id, user1 as friend_id,\n",
        "          datetime\n",
        "          From user_month\n",
        "        ),\n",
        "        temp2 as (\n",
        "          Select user_id, min(datetime) as firstDate\n",
        "          From temp1\n",
        "          group by user_id\n",
        "        ),\n",
        "        temp3 as (\n",
        "          Select temp1.user_id as user_id, friend_id, datetime, firstDate,\n",
        "          DATEDIFF(month, firstDate, datetime) as month_num,\n",
        "          dense_rank() OVER(PARTITION BY temp1.user_id, friend_id ORDER BY datetime) AS rank_num\n",
        "          From temp1\n",
        "          Left Join temp2 on temp1.user_id = temp2.user_id\n",
        "          where DATEDIFF(month, firstDate, datetime) between 0 and 12\n",
        "        ),\n",
        "        temp4 as (\n",
        "          Select distinct user_id, friend_id, month_num, 1 As Contact_Num\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "          UNION\n",
        "          Select distinct A.user_id as user_id,\n",
        "          B.friend_id as friend_id,\n",
        "          DATEDIFF(month, B.datetime, A.firstDate) as month_num,\n",
        "          2 AS Contact_Num\n",
        "          From temp3 A\n",
        "          Left JOIN temp3 B on A.friend_id = B.user_id\n",
        "          and A.user_id <> B.friend_id\n",
        "          where B.friend_id is not NULL\n",
        "          and DATEDIFF(month, B.datetime, A.firstDate) between 0 and 12\n",
        "        ),\n",
        "        temp5 as (\n",
        "          Select *,\n",
        "          row_number() OVER(PARTITION BY user_id, friend_id, month_num ORDER BY Contact_Num) AS rank_num\n",
        "          FROM temp4\n",
        "        ),\n",
        "        temp6 as (\n",
        "          Select user_id, month_num, friend_id as friends_friends\n",
        "          from temp5\n",
        "          where rank_num=1 and Contact_Num = 2\n",
        "        ),\n",
        "        temp7 as (\n",
        "          Select distinct user_id,\n",
        "          month_num,\n",
        "          friend_id as friends\n",
        "          From temp3\n",
        "          where rank_num =1\n",
        "          ),\n",
        "        temp8 as (\n",
        "          Select dummy_month.user_id, dummy_month.month_num\n",
        "          ,friends\n",
        "          ,friends_friends\n",
        "          From dummy_month\n",
        "          Left JOIN temp7 on temp7.user_id = dummy_month.user_id and temp7.month_num = dummy_month.month_num\n",
        "          Left JOIN temp6 ON temp6.user_id = dummy_month.user_id and temp6.month_num = dummy_month.month_num\n",
        "          )\n",
        "          Select temp8.user_id as user_id, temp8.month_num as month_num\n",
        "        ,collect_list(friends) OVER(PARTITION BY user_id order by month_num) AS friends_list\n",
        "        ,collect_list(friends_friends) OVER(PARTITION BY user_id order by month_num) AS friends_friends_list\n",
        "        from temp8\n",
        "        order by user_id, month_num\n",
        "\n",
        "\"\"\"\n",
        "\n",
        "df_f_fof_list = spark.sql(query)\n",
        "df_f_fof_list.createOrReplaceTempView(\"df_f_fof_list\")\n",
        "df_f_fof_list.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z8YdOc0tgeyo",
        "outputId": "32c09d92-9103-4842-b727-a93bdb1b528e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+------------+--------------------+\n",
            "|user_id|month_num|friends_list|friends_friends_list|\n",
            "+-------+---------+------------+--------------------+\n",
            "|     10|        0|     [71056]|                  []|\n",
            "|     10|        1|     [71056]|                  []|\n",
            "|     10|        2|     [71056]|                  []|\n",
            "|     10|        3|     [71056]|                  []|\n",
            "|     10|        4|     [71056]|                  []|\n",
            "|     10|        5|     [71056]|                  []|\n",
            "|     10|        6|     [71056]|             [47104]|\n",
            "|     10|        7|     [71056]|             [47104]|\n",
            "|     10|        8|     [71056]|             [47104]|\n",
            "|     10|        9|     [71056]|             [47104]|\n",
            "|     10|       10|     [71056]|             [47104]|\n",
            "|     10|       11|     [71056]|             [47104]|\n",
            "|     10|       12| [71056, 43]|             [47104]|\n",
            "|1000013|        0|    [882320]|                  []|\n",
            "|1000013|        1|    [882320]|                  []|\n",
            "|1000013|        2|    [882320]|                  []|\n",
            "|1000013|        3|    [882320]|                  []|\n",
            "|1000013|        4|    [882320]|                  []|\n",
            "|1000013|        5|    [882320]|                  []|\n",
            "|1000013|        6|    [882320]|                  []|\n",
            "+-------+---------+------------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JRekvkXByRaY"
      },
      "source": [
        "### Clustering coefficient of a user's network"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.window import Window"
      ],
      "metadata": {
        "id": "Oy80w29GBB7o"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define a window specification to partition the data by user_id and order by month\n",
        "windowSpec = Window.partitionBy(\"user_id\").orderBy(\"month_num\")"
      ],
      "metadata": {
        "id": "EWZNSqpSbwwj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Calculate the clustering coefficient month-wise\n",
        "df_f_fof = df_f_fof.withColumn(\"triangles\", (df_f_fof[\"friends_count_cum\"] * (df_f_fof[\"friends_count_cum\"] - 1)) / 2)\n",
        "df_f_fof.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IUg-Qe0ebzob",
        "outputId": "2c438377-4fbc-4f46-f38f-9a49240b5d30"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+---------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|triangles|\n",
            "+-------+---------+-----------------+-------------------------+---------+\n",
            "|     10|        0|                1|                        0|      0.0|\n",
            "|     10|        1|                1|                        0|      0.0|\n",
            "|     10|        2|                1|                        0|      0.0|\n",
            "|     10|        3|                1|                        0|      0.0|\n",
            "|     10|        4|                1|                        0|      0.0|\n",
            "|     10|        5|                1|                        0|      0.0|\n",
            "|     10|        6|                1|                        1|      0.0|\n",
            "|     10|        7|                1|                        1|      0.0|\n",
            "|     10|        8|                1|                        1|      0.0|\n",
            "|     10|        9|                1|                        1|      0.0|\n",
            "|     10|       10|                1|                        1|      0.0|\n",
            "|     10|       11|                1|                        1|      0.0|\n",
            "|     10|       12|                2|                        1|      1.0|\n",
            "|1000013|        0|                1|                        0|      0.0|\n",
            "|1000013|        1|                1|                        0|      0.0|\n",
            "|1000013|        2|                1|                        0|      0.0|\n",
            "|1000013|        3|                1|                        0|      0.0|\n",
            "|1000013|        4|                1|                        0|      0.0|\n",
            "|1000013|        5|                1|                        0|      0.0|\n",
            "|1000013|        6|                1|                        0|      0.0|\n",
            "+-------+---------+-----------------+-------------------------+---------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_f_fof = df_f_fof.withColumn(\"possible_triangles\", (df_f_fof[\"friends_count_cum\"] + df_f_fof[\"friends_friends_count_cum\"]) * (df_f_fof[\"friends_count_cum\"] + df_f_fof[\"friends_friends_count_cum\"] - 1) / 2)\n",
        "df_f_fof.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "REpt8la4cw3x",
        "outputId": "1a9c4c5c-dc8c-495e-eb33-e32cad930168"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+---------+------------------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|triangles|possible_triangles|\n",
            "+-------+---------+-----------------+-------------------------+---------+------------------+\n",
            "|     10|        0|                1|                        0|      0.0|               0.0|\n",
            "|     10|        1|                1|                        0|      0.0|               0.0|\n",
            "|     10|        2|                1|                        0|      0.0|               0.0|\n",
            "|     10|        3|                1|                        0|      0.0|               0.0|\n",
            "|     10|        4|                1|                        0|      0.0|               0.0|\n",
            "|     10|        5|                1|                        0|      0.0|               0.0|\n",
            "|     10|        6|                1|                        1|      0.0|               1.0|\n",
            "|     10|        7|                1|                        1|      0.0|               1.0|\n",
            "|     10|        8|                1|                        1|      0.0|               1.0|\n",
            "|     10|        9|                1|                        1|      0.0|               1.0|\n",
            "|     10|       10|                1|                        1|      0.0|               1.0|\n",
            "|     10|       11|                1|                        1|      0.0|               1.0|\n",
            "|     10|       12|                2|                        1|      1.0|               3.0|\n",
            "|1000013|        0|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        1|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        2|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        3|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        4|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        5|                1|                        0|      0.0|               0.0|\n",
            "|1000013|        6|                1|                        0|      0.0|               0.0|\n",
            "+-------+---------+-----------------+-------------------------+---------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_f_fof = df_f_fof.withColumn(\"clustering_coefficient\", F.when(df_f_fof[\"possible_triangles\"] > 0, df_f_fof[\"triangles\"] / df_f_fof[\"possible_triangles\"]).otherwise(0))\n",
        "df_f_fof.createOrReplaceTempView(\"df_f_fof\")\n",
        "df_f_fof.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V06XWbXqdfnX",
        "outputId": "ed865d17-3c3a-49d9-cd47-67fb36486a9a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+---------+------------------+----------------------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|triangles|possible_triangles|clustering_coefficient|\n",
            "+-------+---------+-----------------+-------------------------+---------+------------------+----------------------+\n",
            "|     10|        0|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        1|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        2|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        3|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        4|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        5|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|     10|        6|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|        7|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|        8|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|        9|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|       10|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|       11|                1|                        1|      0.0|               1.0|                   0.0|\n",
            "|     10|       12|                2|                        1|      1.0|               3.0|    0.3333333333333333|\n",
            "|1000013|        0|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        1|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        2|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        3|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        4|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        5|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "|1000013|        6|                1|                        0|      0.0|               0.0|                   0.0|\n",
            "+-------+---------+-----------------+-------------------------+---------+------------------+----------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DVRSV2fJxHoQ"
      },
      "source": [
        "### Page rank of each user"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from graphframes import GraphFrame"
      ],
      "metadata": {
        "id": "nEmLsL-dxCIQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "vertices = user_friends.selectExpr(\"user_id as id\").distinct()\n",
        "edges = user_friends.selectExpr(\"user_id as src\", \"friend_id as dst\")\n",
        "graph = GraphFrame(vertices, edges)\n",
        "pagerank = graph.pageRank(resetProbability=0.15, tol=0.01)\n",
        "df_pagerank = pagerank.vertices.selectExpr(\"id as user_id\", \"pagerank\").orderBy(\"pagerank\", ascending=False)\n",
        "df_pagerank.createOrReplaceTempView(\"df_pagerank\")\n",
        "df_pagerank.show()"
      ],
      "metadata": {
        "id": "2qNOowP8yM6Q",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "581c939e-c433-412e-c2d2-5fe69ec976a9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/pyspark/sql/dataframe.py:169: UserWarning: DataFrame.sql_ctx is an internal property, and will be removed in future releases. Use DataFrame.sparkSession instead.\n",
            "  warnings.warn(\n",
            "/usr/local/lib/python3.10/dist-packages/pyspark/sql/dataframe.py:148: UserWarning: DataFrame constructor is internal. Do not directly use it.\n",
            "  warnings.warn(\"DataFrame constructor is internal. Do not directly use it.\")\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+------------------+\n",
            "|user_id|          pagerank|\n",
            "+-------+------------------+\n",
            "|1002607| 9.569389128569794|\n",
            "| 820531| 6.602097862209916|\n",
            "|6671483|2.7174323451866407|\n",
            "|1116467|2.3004584985886547|\n",
            "| 170573|2.3004584985886547|\n",
            "|2076699|2.3004584985886547|\n",
            "| 351821|2.3004584985886547|\n",
            "|1265238|2.3004584985886547|\n",
            "| 724531|2.3004584985886547|\n",
            "| 354930|2.3004584985886547|\n",
            "| 759712|2.3004584985886547|\n",
            "| 248224|2.2209268318494257|\n",
            "| 831087|2.1949498763871733|\n",
            "| 433480| 2.194949876387173|\n",
            "| 537760| 2.194949876387173|\n",
            "|2581403|2.1047283031943644|\n",
            "|1227346|2.1047283031943644|\n",
            "| 613769|2.0391985842582367|\n",
            "|1738566|2.0328780183573394|\n",
            "| 185517|  2.01448310594115|\n",
            "+-------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Final SNA output"
      ],
      "metadata": {
        "id": "CeN_LNp4IIrk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Count friends of friends for all users\n",
        "query = \"\"\"\n",
        "        Select A.user_id, A.month_num\n",
        "        ,friends_count_cum\n",
        "        ,friends_friends_count_cum\n",
        "        ,clustering_coefficient\n",
        "        ,pagerank\n",
        "        from df_f_fof A\n",
        "        Left Join df_pagerank B on A.user_id = B.user_id\n",
        "        order by user_id, month_num\n",
        "\"\"\"\n",
        "\n",
        "df_venmo_sna = spark.sql(query)\n",
        "df_venmo_sna.createOrReplaceTempView(\"df_venmo_sna\")\n",
        "df_venmo_sna.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9H2fndUtALx2",
        "outputId": "030f240a-f484-4b90-9acb-c4ffa21dc673"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|clustering_coefficient|          pagerank|\n",
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "|     10|        0|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        1|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        2|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        3|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        4|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        5|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        6|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        7|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        8|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        9|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       10|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       11|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       12|                2|                        1|    0.3333333333333333|1.2772318716875664|\n",
            "|1000013|        0|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        1|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        2|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        3|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        4|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        5|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        6|                1|                        0|                   0.0|1.0016935844938735|\n",
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# df_venmo_sna.write.csv(\"venmo_sna.csv\", header=True, mode=\"overwrite\")"
      ],
      "metadata": {
        "id": "hM9HQbo0uIgU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Predictive Analytics with MLlib"
      ],
      "metadata": {
        "id": "GhQz_IPwRpYk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark = SparkSession\\\n",
        "        .builder\\\n",
        "        .getOrCreate()"
      ],
      "metadata": {
        "id": "X0uRQrwiSFma"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### How many transactions user had committed during her/his twelve months in Venmo? (Y variable for the predictive analytics)"
      ],
      "metadata": {
        "id": "KbaNhxaapIQu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# SparkSession setup\n",
        "spark = SparkSession.builder.appName(\"Venmo\").getOrCreate()\n",
        "\n",
        "# Loading the dataset\n",
        "dataframe = spark.read.parquet(\"VenmoSample.snappy.parquet\")\n",
        "# Isolating unique user1 IDs\n",
        "distinct_user1_ids = dataframe.select(\"user1\").distinct()\n",
        "\n",
        "# Subsetting 10% of the user1 IDs randomly\n",
        "random_subset_ids = distinct_user1_ids.sample(withReplacement=False, fraction=0.01, seed=11111)\n",
        "\n",
        "random_subset_ids.show()\n",
        "\n",
        "# Filtering the original dataset with the subset of user1 IDs\n",
        "filtered_dataframe = dataframe.join(random_subset_ids, on=\"user1\")\n",
        "\n",
        "filtered_dataframe.show()\n",
        "\n",
        "filtered_dataframe.createOrReplaceTempView(\"Sample\")\n",
        "\n",
        "from pyspark.sql import functions as F\n",
        "\n",
        "# Calculate max and min dates for each user\n",
        "date_ranges = filtered_dataframe.groupBy(\"user1\").agg(\n",
        "    F.max(\"datetime\").alias(\"last_transaction\"),\n",
        "    F.min(\"datetime\").alias(\"first_transaction\")\n",
        ")\n",
        "\n",
        "# Filter users who have been on the platform for over a year\n",
        "valid_users = date_ranges.filter(\n",
        "    F.datediff(F.col(\"last_transaction\"), F.col(\"first_transaction\")) > 365\n",
        ")\n",
        "\n",
        "# Join back to the original dataframe to retain transaction records along with last_transaction\n",
        "valid_transactions = filtered_dataframe.join(\n",
        "    valid_users,\n",
        "    on=\"user1\"\n",
        ")\n",
        "\n",
        "# Calculate the total number of transactions in the first year for each user\n",
        "transactions_count = valid_transactions.filter(\n",
        "    (F.col(\"datetime\") >= F.date_sub(F.col(\"last_transaction\"), 365)) &\n",
        "    (F.col(\"datetime\") <= F.col(\"last_transaction\"))\n",
        ").groupBy(\"user1\").agg(\n",
        "    F.count(\"datetime\").alias(\"total_transactions\")\n",
        ")\n",
        "\n",
        "transactions_count.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Oosj4yFWpFNi",
        "outputId": "863acbf8-3ba0-4c3c-a650-a3c94e2faaf6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+\n",
            "|  user1|\n",
            "+-------+\n",
            "|   6620|\n",
            "| 243392|\n",
            "| 317828|\n",
            "| 321666|\n",
            "| 333479|\n",
            "| 619248|\n",
            "| 663961|\n",
            "| 707930|\n",
            "| 731446|\n",
            "| 750639|\n",
            "| 751485|\n",
            "| 768789|\n",
            "| 900004|\n",
            "| 934818|\n",
            "| 971791|\n",
            "|1056442|\n",
            "|1074192|\n",
            "|1162591|\n",
            "|1285756|\n",
            "|1330274|\n",
            "+-------+\n",
            "only showing top 20 rows\n",
            "\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "|  user1|  user2|transaction_type|           datetime|         description|is_business|            story_id|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "|1582311| 880966|         payment|2016-05-19 21:49:27|          🏰🌠🎠🎆🚗|      false|573dd277cd03c9af2...|\n",
            "|1136626|3288280|          charge|2016-07-17 19:34:52|            THE DOME|      false|578b7b6c23e064eac...|\n",
            "|2383212|2045295|         payment|2016-01-31 05:08:16|                  🍺|      false|56ad2640cd03c9af2...|\n",
            "|2257760|1933300|         payment|2016-02-06 05:24:19|                  🐭|      false|56b51303cd03c9af2...|\n",
            "|6477174|3529055|         payment|2016-05-23 08:15:37|                🎉🍽|      false|574259b9cd03c9af2...|\n",
            "| 387392| 535788|         payment|2014-12-06 04:47:27|              🐔🐥🐣|      false|548219dffe5f22c0b...|\n",
            "|5397688|2825242|         payment|2015-08-04 08:11:57|         Fried balls|      false|55c0115e5d6cc819a...|\n",
            "|2652941|2008848|         payment|2015-12-01 00:55:56|           Paulieeee|      false|565c7f9ccd03c9af2...|\n",
            "| 943545|2417306|         payment|2016-05-06 07:00:47|                🌮🍷|      false|572bdeafcd03c9af2...|\n",
            "|1016167|3378360|         payment|2016-09-18 14:31:23|            Belmonts|      false|57de42cb23e064eac...|\n",
            "|5396538|8790036|         payment|2016-07-22 08:19:21|                🏠💸|      false|5791749923e064eac...|\n",
            "|2068886|8986632|         payment|2016-05-04 22:59:12|   Installs Ad!!! 😎|      false|572a1c50cd03c9af2...|\n",
            "|2908182|7468842|          charge|2016-04-24 00:48:57|   Brudderhood event|      false|571bb589cd03c9af2...|\n",
            "|4505943|2232284|         payment|2015-08-18 07:16:18| Chinese and beer 🍃|      false|55d27952583c0412f...|\n",
            "|1319361|1404004|         payment|2016-05-25 08:22:39|                🍵🍕|      false|5744fe5fcd03c9af2...|\n",
            "| 332591| 180630|          charge|2014-11-14 02:24:09|Cable Gas and Ele...|      false|5464f7497d0b0354e...|\n",
            "|6224071|1025926|         payment|2016-03-04 14:14:58|              ☀️🌊🎉|      false|56d927e2cd03c9af2...|\n",
            "| 346145|1238063|         payment|2014-11-27 16:39:54|                  🍗|      false|5476e35a0cfa78941...|\n",
            "|  59891| 144581|          charge|2013-08-26 03:46:07|   for bike headlamp|      false|521a6d0fd56b6bac5...|\n",
            "|2906690|5005360|         payment|2015-07-27 06:24:32|       Gas &7-Eleven|      false|55b56c301a3b582ca...|\n",
            "+-------+-------+----------------+-------------------+--------------------+-----------+--------------------+\n",
            "only showing top 20 rows\n",
            "\n",
            "+-------+------------------+\n",
            "|  user1|total_transactions|\n",
            "+-------+------------------+\n",
            "|4207530|                 6|\n",
            "|2262896|                17|\n",
            "|1677250|                 5|\n",
            "|1849772|                 5|\n",
            "|2232389|                 4|\n",
            "|2318886|                 6|\n",
            "|3443754|                13|\n",
            "|5841051|                 5|\n",
            "|1588738|                10|\n",
            "| 220404|                 1|\n",
            "|1253375|                 2|\n",
            "| 753766|                 5|\n",
            "|2554348|                 7|\n",
            "|1972255|                 4|\n",
            "|3932144|                 2|\n",
            "|2705642|                 2|\n",
            "| 618678|                12|\n",
            "| 809364|                 2|\n",
            "| 533484|                 1|\n",
            "| 846231|                 1|\n",
            "+-------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Creating recency and frequency variables for CRM RFM framework. Here, we don’t have monetary amounts, so we will focus on just RF.\n",
        "\n",
        "Recency refers to the last time a user was active, and frequency is how often a user uses Venmo in a month."
      ],
      "metadata": {
        "id": "vKeYwMyrpS4a"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import functions as F\n",
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col\n",
        "from pyspark.sql.window import Window  # Importing the Window class\n",
        "\n",
        "# Convert datetime to date\n",
        "valid_transactions = valid_transactions.withColumn('transaction_date', F.to_date('datetime'))\n",
        "\n",
        "# Define window partitioned by user1\n",
        "user_window = Window.partitionBy('user1')\n",
        "\n",
        "# Find the most recent transaction date for each user and calculate the difference in days and months\n",
        "valid_transactions = valid_transactions.withColumn('latest_transaction_date', F.max('transaction_date').over(user_window))\n",
        "valid_transactions = valid_transactions.withColumn('days_since_transaction', F.datediff('latest_transaction_date', 'transaction_date'))\n",
        "valid_transactions = valid_transactions.withColumn('lifetime_point', (valid_transactions['days_since_transaction'] / 30).cast('integer'))\n",
        "\n",
        "# Filter out transactions that occurred after the first 12 months\n",
        "valid_transactions = valid_transactions.filter(valid_transactions.lifetime_point <= 11)\n",
        "\n",
        "# Group transactions by user1 and months_since_transaction\n",
        "grouped_transactions = valid_transactions.groupBy('user1', 'lifetime_point')\n",
        "\n",
        "# Calculate recency and frequency for each group\n",
        "recency_and_frequency = grouped_transactions.agg(\n",
        "    # Recency is calculated as the maximum number of days in a month (30) minus the maximum number of days since a transaction in that month\n",
        "    (30 * (F.col('lifetime_point') + 1) - F.max('days_since_transaction')).alias('recency'),\n",
        "\n",
        "    # Frequency is calculated as the number of days in a month (30) divided by the number of transactions in that month\n",
        "    (30 / F.count('*')).alias('frequency')\n",
        ")\n",
        "\n",
        "recency_and_frequency.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "koCfblPppPH0",
        "outputId": "a99ccfb6-9561-450d-f98c-910dabf69b8e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------+---------+\n",
            "|user1|lifetime_point|recency|frequency|\n",
            "+-----+--------------+-------+---------+\n",
            "| 5232|             7|     13|     15.0|\n",
            "| 5232|             9|      3|     30.0|\n",
            "| 5232|             0|     30|     30.0|\n",
            "| 5232|             8|     10|     30.0|\n",
            "|36513|             0|     30|     30.0|\n",
            "|46520|             0|     12|     10.0|\n",
            "|46520|             4|      2|     10.0|\n",
            "|46520|             3|     28|     15.0|\n",
            "|46520|             7|      3|     10.0|\n",
            "|46520|             6|     27|     30.0|\n",
            "|49505|             0|      5|     15.0|\n",
            "|49577|             7|      9|     15.0|\n",
            "|49577|             5|     14|     30.0|\n",
            "|49577|             0|     30|     30.0|\n",
            "|52525|             0|     30|     30.0|\n",
            "|62177|             0|     30|     30.0|\n",
            "|65802|            11|      7|     30.0|\n",
            "|65802|             6|     16|     30.0|\n",
            "|65802|             1|     11|     15.0|\n",
            "|65802|             0|     14|     15.0|\n",
            "+-----+--------------+-------+---------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Regress recency and frequency on Total Transactions"
      ],
      "metadata": {
        "id": "XEWq0goXpd8u"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Join the two dataframes\n",
        "combined_df = recency_and_frequency.join(transactions_count, on='user1', how='inner')\n",
        "\n",
        "# Show the combined dataframe\n",
        "combined_df.show()\n",
        "\n",
        "combined_df_single_partition = combined_df.coalesce(1)\n",
        "\n",
        "# Save the combined dataframe as a single CSV file\n",
        "combined_df_single_partition.write.csv(\"combined_df.csv\", header=True, mode=\"overwrite\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DlS9xmM0piUL",
        "outputId": "b2ef4939-2b76-4a04-9f3d-17748d624921"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------+---------+------------------+\n",
            "|user1|lifetime_point|recency|frequency|total_transactions|\n",
            "+-----+--------------+-------+---------+------------------+\n",
            "| 5232|             7|     13|     15.0|                 5|\n",
            "| 5232|             9|      3|     30.0|                 5|\n",
            "| 5232|             0|     30|     30.0|                 5|\n",
            "| 5232|             8|     10|     30.0|                 5|\n",
            "|36513|             0|     30|     30.0|                 1|\n",
            "|46520|             0|     12|     10.0|                12|\n",
            "|46520|             4|      2|     10.0|                12|\n",
            "|46520|             3|     28|     15.0|                12|\n",
            "|46520|             7|      3|     10.0|                12|\n",
            "|46520|             6|     27|     30.0|                12|\n",
            "|49505|             0|      5|     15.0|                 2|\n",
            "|49577|             7|      9|     15.0|                 4|\n",
            "|49577|             5|     14|     30.0|                 4|\n",
            "|49577|             0|     30|     30.0|                 4|\n",
            "|52525|             0|     30|     30.0|                 1|\n",
            "|62177|             0|     30|     30.0|                 1|\n",
            "|65802|            11|      7|     30.0|                 7|\n",
            "|65802|             6|     16|     30.0|                 7|\n",
            "|65802|             1|     11|     15.0|                 7|\n",
            "|65802|             0|     14|     15.0|                 7|\n",
            "+-----+--------------+-------+---------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.ml.regression import LinearRegression\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "mse_dict = {}\n",
        "\n",
        "for i in range(13):\n",
        "    filtered_df = combined_df.filter(combined_df.lifetime_point == i)\n",
        "\n",
        "    if filtered_df.count() == 0:\n",
        "        continue\n",
        "\n",
        "    train_data, test_data = filtered_df.randomSplit([0.8, 0.2], seed=42)\n",
        "\n",
        "    assembler = VectorAssembler(inputCols=[\"recency\", \"frequency\"], outputCol=\"features\")\n",
        "\n",
        "    train_data = assembler.transform(train_data)\n",
        "    test_data = assembler.transform(test_data)\n",
        "\n",
        "    train_dataset = train_data.select(\"features\", \"total_transactions\").withColumnRenamed(\"total_transactions\", \"Y\")\n",
        "    test_dataset = test_data.select(\"features\", \"total_transactions\").withColumnRenamed(\"total_transactions\", \"Y\")\n",
        "\n",
        "    lr = LinearRegression(featuresCol=\"features\", labelCol=\"Y\")\n",
        "\n",
        "    lr_model = lr.fit(train_dataset)\n",
        "\n",
        "    predictions = lr_model.transform(test_dataset)\n",
        "\n",
        "    mse = predictions.selectExpr(\"POWER((Y - prediction), 2) as mse\").groupBy().mean(\"mse\").collect()[0][0]\n",
        "    mse_dict[i] = mse\n",
        "\n",
        "plt.figure(figsize=(8, 6))\n",
        "plt.plot(range(12), mse_dict.values(), marker='o')\n",
        "plt.xlabel(\"Lifetime Point\")\n",
        "plt.ylabel(\"MSE\")\n",
        "plt.title(\"MSE for Each Lifetime Point\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "RTAa1W_-pvGZ",
        "outputId": "b8eb9851-9526-4832-a1a1-49936db314f0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The predictive accuracy of recency and frequency in determining the total number of transactions improves and stabilizes over time, but there may be other influential factors during the early stages that contribute to the variability in the predictions.\n"
      ],
      "metadata": {
        "id": "d8sMxlTx9DDM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Regress recency, frequency AND the spending behavior profile on total transations."
      ],
      "metadata": {
        "id": "_fI878havJkY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession\n",
        "\n",
        "spark = SparkSession.builder.appName('loadData').getOrCreate()\n",
        "\n",
        "file_path = 'Customer behvaior.csv'\n",
        "customer_behavior = spark.read.csv(file_path, header=True)\n",
        "\n",
        "# To see the first few rows of the loaded data\n",
        "customer_behavior.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZGysZo4evINc",
        "outputId": "fc7df157-a7f5-4677-8da8-e4bcc87680b0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------+-----+--------+----+-----+-----+--------------+------+--------------+------+-------+------------------+\n",
            "|   user1|month|Activity|Cash|Event| Food|IllegalSarcasm|People|Transportation|Travel|Utility|dynamic_percentage|\n",
            "+--------+-----+--------+----+-----+-----+--------------+------+--------------+------+-------+------------------+\n",
            "| 1000013|    6|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1000057|    5|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "|10000940|    4|   100.0|   0|    0|    0|             0|     0|             0|     0|      0|             100.0|\n",
            "|10003651|    4|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "| 1000387|    1|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "| 1000501|   10|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "| 1000535|   10|       0|   0|    0|    0|             0|     0|         100.0|     0|      0|             100.0|\n",
            "|10005784|    6|       0|   0|    0|    0|         100.0|     0|             0|     0|      0|             100.0|\n",
            "| 1000628|    8|   100.0|   0|    0|    0|             0|     0|             0|     0|      0|             100.0|\n",
            "|10007866|    7|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1000864|    4|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1001157|   11|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "| 1001184|    7|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "|10013427|    4|   100.0|   0|    0|    0|             0|     0|             0|     0|      0|             100.0|\n",
            "|10017665|    4|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|             100.0|\n",
            "|10019023|   10|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1002102|    5|       0|   0|100.0|    0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1002240|    6|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "|10023521|    8|       0|   0|    0|100.0|             0|     0|             0|     0|      0|             100.0|\n",
            "| 1002514|    6|       0|   0|    0|    0|             0| 100.0|             0|     0|      0|             100.0|\n",
            "+--------+-----+--------+----+-----+-----+--------------+------+--------------+------+-------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession\n",
        "from pyspark.sql.functions import col\n",
        "customer_behavior.createOrReplaceTempView(\"customer_behavior\")\n",
        "recency_and_frequency.createOrReplaceTempView(\"recency_and_frequency\")\n",
        "transactions_count.createOrReplaceTempView(\"transactions_count\")\n",
        "df_regression2 = spark.sql(\"\"\"\n",
        "select transactions_count.user1, lifetime_point, recency, frequency, Activity, Cash, Event, Food, IllegalSarcasm, People, Transportation, Travel, Utility, total_transactions as Y\n",
        "from customer_behavior\n",
        "join transactions_count on customer_behavior.user1 = transactions_count.user1\n",
        "join recency_and_frequency on recency_and_frequency.user1 = customer_behavior.user1\"\"\")\n",
        "df_regression2.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UbU4poPnvewc",
        "outputId": "9e9b3151-3182-490e-a229-fab9bffb8693"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|user1|lifetime_point|recency|frequency|Activity|Cash|Event| Food|IllegalSarcasm|People|Transportation|Travel|Utility|  Y|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|61428|             9|      6|     15.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|            11|     14|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             7|     20|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             4|     18|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             0|     30|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|65258|            11|     19|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             5|     24|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             2|     27|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             6|      3|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             7|      7|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             3|     22|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65523|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             1|      1|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             5|     29|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|71388|             0|     11|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             1|      8|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|            10|     12|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             9|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             8|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# df_regression2.write.csv('combined_regression', header=True, mode='overwrite')"
      ],
      "metadata": {
        "id": "1k3J9Xc2vZqK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#reload data\n",
        "file_path = 'combined_regression'\n",
        "customer_behavior_2 = spark.read.csv(file_path, header=True)\n",
        "\n",
        "# To see the first few rows of the loaded data\n",
        "customer_behavior_2.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wAFav4eHvtA5",
        "outputId": "bfeea52c-ff24-4d9b-82ab-d1fa9478bdea"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|user1|lifetime_point|recency|frequency|Activity|Cash|Event| Food|IllegalSarcasm|People|Transportation|Travel|Utility|  Y|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|61428|             9|      6|     15.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|            11|     14|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             7|     20|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             4|     18|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             0|     30|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|65258|            11|     19|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             5|     24|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             2|     27|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             6|      3|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             7|      7|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             3|     22|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65523|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             1|      1|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             5|     29|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|71388|             0|     11|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             1|      8|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|            10|     12|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             9|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             8|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import col\n",
        "from pyspark.sql.types import DoubleType\n",
        "from pyspark.ml.regression import LinearRegression\n",
        "from pyspark.ml.feature import VectorAssembler\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Convert the \"Y\" column to DoubleType\n",
        "customer_behavior_2 = customer_behavior_2.withColumn(\"Y\", col(\"Y\").cast(DoubleType()))\n",
        "\n",
        "mse_dict = {}\n",
        "\n",
        "for i in range(13):\n",
        "    filtered_df = customer_behavior_2.filter(customer_behavior_2.lifetime_point == i)\n",
        "\n",
        "    if filtered_df.count() == 0:\n",
        "        continue\n",
        "\n",
        "    # Convert the columns to DoubleType\n",
        "    columns_to_cast = [\"recency\", \"frequency\", \"Activity\", \"Cash\",\n",
        "                       \"Event\", \"Food\", \"IllegalSarcasm\", \"People\",\n",
        "                       \"Transportation\", \"Travel\", \"Utility\"]\n",
        "\n",
        "    for col_name in columns_to_cast:\n",
        "        filtered_df = filtered_df.withColumn(col_name, col(col_name).cast(\"double\"))\n",
        "\n",
        "    assembler = VectorAssembler(inputCols=columns_to_cast, outputCol=\"features\")\n",
        "\n",
        "    filtered_df = assembler.transform(filtered_df)\n",
        "\n",
        "    train_data, test_data = filtered_df.randomSplit([0.8, 0.2], seed=42)\n",
        "\n",
        "    train_dataset = train_data.select(\"features\", \"Y\")\n",
        "    test_dataset = test_data.select(\"features\", \"Y\")\n",
        "\n",
        "    lr = LinearRegression(featuresCol=\"features\", labelCol=\"Y\")\n",
        "\n",
        "    lr_model = lr.fit(train_dataset)\n",
        "\n",
        "    predictions = lr_model.transform(test_dataset)\n",
        "\n",
        "    mse = predictions.selectExpr(\"POWER((Y - prediction), 2) as mse\").groupBy().mean(\"mse\").collect()[0][0]\n",
        "    mse_dict[i] = mse\n",
        "\n",
        "# Plot the MSE for each lifetime point\n",
        "plt.figure(figsize=(8, 6))\n",
        "plt.plot(list(mse_dict.keys()), list(mse_dict.values()), marker='o')\n",
        "plt.xlabel(\"Lifetime Point\")\n",
        "plt.ylabel(\"MSE\")\n",
        "plt.title(\"MSE for Each Lifetime Point with Behavior Metrics\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "yK39Tj_Evy5n",
        "outputId": "09005a22-f570-4f27-ff4e-f716adc367f5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Regress the social network metrics on total transactions."
      ],
      "metadata": {
        "id": "x_woj5XqwFkR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "socialnetworkmetrics = spark.read.csv(\"venmo_sna.csv\", header=True)\n",
        "socialnetworkmetrics.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Px_AmCoPwE-g",
        "outputId": "ec002599-7949-4d38-be6a-255a1318b09e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "|user_id|month_num|friends_count_cum|friends_friends_count_cum|clustering_coefficient|          pagerank|\n",
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "|     10|        0|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        1|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        2|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        3|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        4|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        5|                1|                        0|                   0.0|1.2772318716875664|\n",
            "|     10|        6|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        7|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        8|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|        9|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       10|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       11|                1|                        1|                   0.0|1.2772318716875664|\n",
            "|     10|       12|                2|                        1|    0.3333333333333333|1.2772318716875664|\n",
            "|1000013|        0|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        1|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        2|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        3|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        4|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        5|                1|                        0|                   0.0|1.0016935844938735|\n",
            "|1000013|        6|                1|                        0|                   0.0|1.0016935844938735|\n",
            "+-------+---------+-----------------+-------------------------+----------------------+------------------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "customer_behavior_2.createOrReplaceTempView(\"customer_behavior_2\")\n",
        "socialnetworkmetrics.createOrReplaceTempView(\"socialnetworkmetrics\")\n",
        "transactions_count.createOrReplaceTempView(\"transactions_count\")\n",
        "\n",
        "df_regression3 = spark.sql(\"\"\"\n",
        "    SELECT lifetime_point, recency, frequency, pagerank, clustering_coefficient, total_transactions AS Y\n",
        "    FROM recency_and_frequency\n",
        "    JOIN transactions_count ON recency_and_frequency.user1 = transactions_count.user1\n",
        "    JOIN socialnetworkmetrics ON recency_and_frequency.user1 = socialnetworkmetrics.user_id\n",
        "\"\"\")\n",
        "df_regression3.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1robC77owJ3s",
        "outputId": "3e1fbd18-d84f-4090-f372-0ae8b722060b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------------+-------+---------+------------------+----------------------+---+\n",
            "|lifetime_point|recency|frequency|          pagerank|clustering_coefficient|  Y|\n",
            "+--------------+-------+---------+------------------+----------------------+---+\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             8|      1|     30.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "|             6|     15|     10.0|1.0016935844938735|                   0.0|  9|\n",
            "+--------------+-------+---------+------------------+----------------------+---+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.ml.feature import VectorAssembler\n",
        "from pyspark.ml.regression import LinearRegression\n",
        "from pyspark.ml.evaluation import RegressionEvaluator\n",
        "from pyspark.sql.functions import col\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Convert string columns to numeric types\n",
        "df_regression3 = df_regression3.withColumn(\"pagerank\", col(\"pagerank\").cast(\"double\"))\n",
        "df_regression3 = df_regression3.withColumn(\"clustering_coefficient\", col(\"clustering_coefficient\").cast(\"double\"))\n",
        "\n",
        "# Prepare the data for regression\n",
        "assembler = VectorAssembler(inputCols=[\"recency\", \"frequency\", \"pagerank\", \"clustering_coefficient\"], outputCol=\"features\")\n",
        "regression_data = assembler.transform(df_regression3)\n",
        "\n",
        "mse_values = []\n",
        "\n",
        "# Perform regression analysis for each lifetime point\n",
        "for point in range(1, 12):\n",
        "    # Filter data for the current lifetime point\n",
        "    point_data = regression_data.filter(regression_data.lifetime_point == point)\n",
        "\n",
        "    # Split the data into training and testing sets\n",
        "    (training_data, testing_data) = point_data.randomSplit([0.7, 0.3])\n",
        "\n",
        "    # Create a Linear Regression model\n",
        "    lr = LinearRegression(featuresCol=\"features\", labelCol=\"Y\")\n",
        "\n",
        "    # Train the model\n",
        "    model = lr.fit(training_data)\n",
        "\n",
        "    # Make predictions on the testing data\n",
        "    predictions = model.transform(testing_data)\n",
        "\n",
        "    # Calculate the MSE\n",
        "    evaluator = RegressionEvaluator(labelCol=\"Y\", predictionCol=\"prediction\", metricName=\"mse\")\n",
        "    mse = evaluator.evaluate(predictions)\n",
        "    mse_values.append(mse)\n",
        "\n",
        "# Plot the MSE for each lifetime point\n",
        "plt.plot(range(1, 12), mse_values, marker='o')\n",
        "plt.xlabel(\"Lifetime Point\")\n",
        "plt.ylabel(\"MSE\")\n",
        "plt.title(\"MSE for Social Network Metrics at Each Lifetime Point\")\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "-dvskLiMwcDj",
        "outputId": "1b348a06-33a1-4f46-e44b-ae99a8e956fa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The graph highlights the varying performance of the prediction model at different stages of the customer lifetime. Similarly, when the social metrics variable is added, the overall trend goes downward but with intermittent increases and decreases."
      ],
      "metadata": {
        "id": "YGjDnEWq9P_l"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Regress the social network metrics and the spending behavior of social network on total transactions."
      ],
      "metadata": {
        "id": "D_5SP5Vaw7gW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#reload data\n",
        "file_path = 'combined_regression'\n",
        "customer_behavior_2 = spark.read.csv(file_path, header=True)\n",
        "\n",
        "# To see the first few rows of the loaded data\n",
        "customer_behavior_2.show()\n",
        "\n",
        "# # To see the first few rows of the loaded data\n",
        "\n",
        "\n",
        "customer_behavior_2.createOrReplaceTempView(\"customer_behavior\")\n",
        "transactions_count.createOrReplaceTempView(\"transactions_count\")\n",
        "socialnetworkmetrics.createOrReplaceTempView(\"socialnetworkmetrics\")\n",
        "\n",
        "df_regression4 = spark.sql(\"\"\"\n",
        "    SELECT `customer_behavior`.`lifetime_point`, `customer_behavior`.`recency`, `customer_behavior`.`frequency`, `socialnetworkmetrics`.`pagerank`, `socialnetworkmetrics`.`clustering_coefficient`, `customer_behavior`.`Activity`, `customer_behavior`.`Cash`, `customer_behavior`.`Event`, `customer_behavior`.`Food`, `customer_behavior`.`IllegalSarcasm`, `customer_behavior`.`People`, `customer_behavior`.`Transportation`, `customer_behavior`.`Travel`, `customer_behavior`.`Utility`, `transactions_count`.`total_transactions` AS Y\n",
        "    FROM socialnetworkmetrics\n",
        "    JOIN transactions_count ON socialnetworkmetrics.user_id = transactions_count.user1\n",
        "    JOIN customer_behavior ON customer_behavior.user1 = socialnetworkmetrics.user_id\n",
        "\"\"\")\n",
        "\n",
        "df_regression4.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1LZLl2g8w8N7",
        "outputId": "bfdc0a6d-9ac7-40a6-b02f-9c0e3e847323"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|user1|lifetime_point|recency|frequency|Activity|Cash|Event| Food|IllegalSarcasm|People|Transportation|Travel|Utility|  Y|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "|61428|             9|      6|     15.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|            11|     14|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             7|     20|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             4|     18|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|61428|             0|     30|     30.0|       0|   0|    0|    0|             0|     0|             0|     0|  100.0|  6|\n",
            "|65258|            11|     19|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             5|     24|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             2|     27|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             6|      3|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             7|      7|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65258|             3|     22|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  8|\n",
            "|65523|             0|     30|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             1|      1|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|65523|             5|     29|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  3|\n",
            "|71388|             0|     11|     15.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             1|      8|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|            10|     12|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             9|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "|71388|             8|      5|     30.0|       0|   0|    0|100.0|             0|     0|             0|     0|      0|  6|\n",
            "+-----+--------------+-------+---------+--------+----+-----+-----+--------------+------+--------------+------+-------+---+\n",
            "only showing top 20 rows\n",
            "\n",
            "+--------------+-------+---------+------------------+----------------------+--------+----+-----+----+--------------+------+--------------+------+-------+---+\n",
            "|lifetime_point|recency|frequency|          pagerank|clustering_coefficient|Activity|Cash|Event|Food|IllegalSarcasm|People|Transportation|Travel|Utility|  Y|\n",
            "+--------------+-------+---------+------------------+----------------------+--------+----+-----+----+--------------+------+--------------+------+-------+---+\n",
            "|             4|      8|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             7|     13|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             1|     11|     10.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             3|      9|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|            11|      2|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             6|      4|     10.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             0|     30|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             8|     18|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             4|      8|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             7|     13|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             1|     11|     10.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             3|      9|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|            11|      2|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             6|      4|     10.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             0|     30|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             8|     18|     30.0|1.0016935844938735|                   0.0|       0|   0|100.0|   0|             0|     0|             0|     0|      0| 13|\n",
            "|             4|      8|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             7|     13|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             1|     11|     10.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "|             3|      9|     30.0|1.0016935844938735|                   0.0|       0|   0|    0|   0|             0|     0|         100.0|     0|      0| 13|\n",
            "+--------------+-------+---------+------------------+----------------------+--------+----+-----+----+--------------+------+--------------+------+-------+---+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Cast columns to the appropriate data types\n",
        "df_regression4 = df_regression4.withColumn(\"pagerank\", col(\"pagerank\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"clustering_coefficient\", col(\"clustering_coefficient\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Activity\", col(\"Activity\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Cash\", col(\"Cash\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Event\", col(\"Event\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Food\", col(\"Food\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"IllegalSarcasm\", col(\"IllegalSarcasm\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"People\", col(\"People\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Transportation\", col(\"Transportation\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Travel\", col(\"Travel\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Utility\", col(\"Utility\").cast(\"double\"))\n",
        "df_regression4 = df_regression4.withColumn(\"Y\", col(\"Y\").cast(\"double\"))\n",
        "\n",
        "# Assemble the features into a vector\n",
        "assembler = VectorAssembler(\n",
        "    inputCols=[\"pagerank\", \"clustering_coefficient\", \"Activity\", \"Cash\", \"Event\", \"Food\", \"IllegalSarcasm\", \"People\", \"Transportation\", \"Travel\", \"Utility\"],\n",
        "    outputCol=\"features\")\n",
        "\n",
        "# Apply the vector assembler on the dataframe\n",
        "df_assembled = assembler.transform(df_regression4)\n",
        "\n",
        "# Split the data into train and test sets\n",
        "train_data, test_data = df_assembled.randomSplit([0.8, 0.2], seed=42)\n",
        "\n",
        "# Initialize the linear regression model\n",
        "lr = LinearRegression(labelCol=\"Y\")\n",
        "\n",
        "# Initialize a dictionary to store the MSE for each lifetime point\n",
        "mse_dict = {}\n",
        "\n",
        "for lifetime_point in range(13):\n",
        "    # Filter the data for the current lifetime point\n",
        "    train_data_filtered = train_data.filter(train_data.lifetime_point == lifetime_point)\n",
        "    test_data_filtered = test_data.filter(test_data.lifetime_point == lifetime_point)\n",
        "\n",
        "    # Check if the training dataset is not empty\n",
        "    if train_data_filtered.count() == 0:\n",
        "        print(f\"No data available for lifetime point {lifetime_point}\")\n",
        "        continue\n",
        "\n",
        "    # Fit the linear regression model on the training data\n",
        "    model = lr.fit(train_data_filtered)\n",
        "\n",
        "    # Make predictions on the test data\n",
        "    predictions = model.transform(test_data_filtered)\n",
        "\n",
        "    # Evaluate the model using MSE\n",
        "    evaluator = RegressionEvaluator(labelCol=\"Y\", metricName=\"mse\")\n",
        "    mse = evaluator.evaluate(predictions)\n",
        "\n",
        "    # Store the MSE for the current lifetime point\n",
        "    mse_dict[lifetime_point] = mse\n",
        "\n",
        "# Extract the lifetime points and MSE values\n",
        "lifetime_points = list(mse_dict.keys())\n",
        "mse_values = list(mse_dict.values())\n",
        "\n",
        "# Plot the MSE for each lifetime point\n",
        "plt.plot(lifetime_points, mse_values, marker='o')\n",
        "plt.xlabel(\"Lifetime Point\")\n",
        "plt.ylabel(\"MSE\")\n",
        "plt.title(\"MSE for Each Lifetime Point\")\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 508
        },
        "id": "T9-6BhY96Zxe",
        "outputId": "16218654-2f9e-4a68-85a1-6cc7f23c9690"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "No data available for lifetime point 9\n",
            "No data available for lifetime point 12\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "When both variables are added, the MSE drops between lifetime 6-8 and then increases drastically."
      ],
      "metadata": {
        "id": "_j5iVjqC9YQu"
      }
    }
  ]
}