car_data_regressao_linear_multipla.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regressão Linear Múltipla"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Objetivos**\n",
    "\n",
    "* Utilizar a biblioteca scikit-learn para construir um modelo de previsão de preço, usando regressão linear múltipla (múltiplas variáveis preditoras que podem ser contínuas ou categóricas). Para isso será utilizado um conjunto de dados de carros usados listados pelo site www.cardekho.com e disponível em https://www.kaggle.com/nehalbirla/vehicle-dataset-from-cardekho e adaptado neste notebook."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Características do Data Set**\n",
    "\n",
    "* **nome_carro:**        nome do carro\n",
    "\n",
    "* **ano:**               ano de fabricação\n",
    "\n",
    "* **preco_venda:**       preço que o dono quer vender\n",
    "\n",
    "* **preco_atual:**       preço da concessionária\n",
    "\n",
    "* **kms_rodados:**       kilometragem do carro \n",
    "\n",
    "* **tipo_combustivel:**  tipo de combustível do carro (Diesel, Gasolina, GasNatural)\n",
    "\n",
    "* **tipo_vendedor:**     se o vendedor é um revendedor ou é a própria pessoa\n",
    "\n",
    "* **tipo_transmissao:**  tipo de câmbio do carro\n",
    "\n",
    "* **n_donos:**           número de donos que o carro já teve\n",
    "\n",
    "\n",
    "* Linhas: 301\n",
    "* Colunas: 9\n",
    "* Formato do arquivo: csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Importando bibliotecas necessárias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "# Line magic function\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Carregando os dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>nome_carro</th>\n",
       "      <th>ano</th>\n",
       "      <th>preco_venda</th>\n",
       "      <th>preco_atual</th>\n",
       "      <th>kms_rodados</th>\n",
       "      <th>tipo_combustivel</th>\n",
       "      <th>tipo_vendedor</th>\n",
       "      <th>tipo_transmissao</th>\n",
       "      <th>n_donos</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ritz</td>\n",
       "      <td>2014</td>\n",
       "      <td>3.35</td>\n",
       "      <td>5.59</td>\n",
       "      <td>27000</td>\n",
       "      <td>Gasolina</td>\n",
       "      <td>Revendedor</td>\n",
       "      <td>Manual</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>sx4</td>\n",
       "      <td>2013</td>\n",
       "      <td>4.75</td>\n",
       "      <td>9.54</td>\n",
       "      <td>43000</td>\n",
       "      <td>Diesel</td>\n",
       "      <td>Revendedor</td>\n",
       "      <td>Manual</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ciaz</td>\n",
       "      <td>2017</td>\n",
       "      <td>7.25</td>\n",
       "      <td>9.85</td>\n",
       "      <td>6900</td>\n",
       "      <td>Gasolina</td>\n",
       "      <td>Revendedor</td>\n",
       "      <td>Manual</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>wagon r</td>\n",
       "      <td>2011</td>\n",
       "      <td>2.85</td>\n",
       "      <td>4.15</td>\n",
       "      <td>5200</td>\n",
       "      <td>Gasolina</td>\n",
       "      <td>Revendedor</td>\n",
       "      <td>Manual</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>swift</td>\n",
       "      <td>2014</td>\n",
       "      <td>4.60</td>\n",
       "      <td>6.87</td>\n",
       "      <td>42450</td>\n",
       "      <td>Diesel</td>\n",
       "      <td>Revendedor</td>\n",
       "      <td>Manual</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  nome_carro   ano  preco_venda  preco_atual  kms_rodados tipo_combustivel  \\\n",
       "0       ritz  2014         3.35         5.59        27000         Gasolina   \n",
       "1        sx4  2013         4.75         9.54        43000           Diesel   \n",
       "2       ciaz  2017         7.25         9.85         6900         Gasolina   \n",
       "3    wagon r  2011         2.85         4.15         5200         Gasolina   \n",
       "4      swift  2014         4.60         6.87        42450           Diesel   \n",
       "\n",
       "  tipo_vendedor tipo_transmissao  n_donos  \n",
       "0    Revendedor           Manual        0  \n",
       "1    Revendedor           Manual        0  \n",
       "2    Revendedor           Manual        0  \n",
       "3    Revendedor           Manual        0  \n",
       "4    Revendedor           Manual        0  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"car_data.csv\")\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(301, 9)\n"
     ]
    }
   ],
   "source": [
    "# O dataset possui 301 registros de carros usados\n",
    "print(df.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para o algoritmo de regressão, substituir os dados que possuem textos por números. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>nome_carro</th>\n",
       "      <th>ano</th>\n",
       "      <th>preco_venda</th>\n",
       "      <th>preco_atual</th>\n",
       "      <th>kms_rodados</th>\n",
       "      <th>tipo_combustivel</th>\n",
       "      <th>tipo_vendedor</th>\n",
       "      <th>tipo_transmissao</th>\n",
       "      <th>n_donos</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ritz</td>\n",
       "      <td>2014</td>\n",
       "      <td>3.35</td>\n",
       "      <td>5.59</td>\n",
       "      <td>27000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>sx4</td>\n",
       "      <td>2013</td>\n",
       "      <td>4.75</td>\n",
       "      <td>9.54</td>\n",
       "      <td>43000</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ciaz</td>\n",
       "      <td>2017</td>\n",
       "      <td>7.25</td>\n",
       "      <td>9.85</td>\n",
       "      <td>6900</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>wagon r</td>\n",
       "      <td>2011</td>\n",
       "      <td>2.85</td>\n",
       "      <td>4.15</td>\n",
       "      <td>5200</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>swift</td>\n",
       "      <td>2014</td>\n",
       "      <td>4.60</td>\n",
       "      <td>6.87</td>\n",
       "      <td>42450</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  nome_carro   ano  preco_venda  preco_atual  kms_rodados  tipo_combustivel  \\\n",
       "0       ritz  2014         3.35         5.59        27000                 0   \n",
       "1        sx4  2013         4.75         9.54        43000                 1   \n",
       "2       ciaz  2017         7.25         9.85         6900                 0   \n",
       "3    wagon r  2011         2.85         4.15         5200                 0   \n",
       "4      swift  2014         4.60         6.87        42450                 1   \n",
       "\n",
       "   tipo_vendedor  tipo_transmissao  n_donos  \n",
       "0              0                 0        0  \n",
       "1              0                 0        0  \n",
       "2              0                 0        0  \n",
       "3              0                 0        0  \n",
       "4              0                 0        0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['tipo_transmissao'].replace({'Manual': 0, 'Automatico': 1}, inplace=True)\n",
    "df['tipo_vendedor'].replace({'Revendedor': 0, 'Individual': 1}, inplace=True)\n",
    "df['tipo_combustivel'].replace({'Gasolina': 0, 'Diesel': 1, 'GasNatural': 2}, inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plotar em um gráfico os anos do carros versus o preço de venda"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(df['ano'], df['preco_venda'],  color='blue')\n",
    "plt.xlabel(\"Ano do Carro\")\n",
    "plt.ylabel(\"Preço de venda\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos ver qual os anos dos carros que mais rodaram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(df['ano'], df['kms_rodados'],  color='red')\n",
    "plt.xlabel(\"Ano do Carro\")\n",
    "plt.ylabel(\"Kms Rodados\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Separando a variável dependente e as variáveis independentes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# escolhendo as variáveis independentes\n",
    "x = df[['ano', 'preco_atual', 'kms_rodados', 'tipo_combustivel', 'tipo_vendedor', 'tipo_transmissao', 'n_donos']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# variável dependente\n",
    "y = df[['preco_venda']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Separar dados em treino e teste"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para construir nosso modelo de Machine Learning precisamos dividir os dados em duas partes. Uma para treinar o modelo\n",
    "e outra para testar o modelo com dados que ele não conhece. Para isso iremos separar 80% dos dados para treino e 20%\n",
    "dos dados para teste. Para isso usaremos a biblioteca train_test_split que faz essa separação automaticamente."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**x_train** é a parte usada para treinar o algoritmo\n",
    "\n",
    "**x_test** é a parte usada para testar o algoritmo depois de treinado\n",
    "\n",
    "**y_train** é a parte das respostas usadas no treino\n",
    "\n",
    "**y_test** é a parte das respostas usadas para testar o algoritmo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Treino\n",
      "x_train:  (240, 7)\n",
      "y_train:  (240, 1)\n",
      "\n",
      "Teste\n",
      "x_test:  (61, 7)\n",
      "y_test:  (61, 1)\n"
     ]
    }
   ],
   "source": [
    "# Vendo o número de registros que serão usados para o treino e teste\n",
    "print('Treino')\n",
    "print(\"x_train: \",x_train.shape)\n",
    "print(\"y_train: \",y_train.shape)\n",
    "\n",
    "print('\\nTeste')\n",
    "print(\"x_test: \",x_test.shape)\n",
    "print(\"y_test: \",y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Treinando o Modelo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "# define um modelo de regressão\n",
    "model = LinearRegression()\n",
    "\n",
    "# treina o modelo com os dados de treino\n",
    "model.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Realiza a previsão dos preços"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# y_pred contém a previsão dos preços do nosso modelo\n",
    "y_pred = model.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(61, 1)\n",
      "(61, 1)\n"
     ]
    }
   ],
   "source": [
    "print(y_pred.shape)\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Agora vamos plotar o preço dos carros, previsto pelo modelo versus os dados de teste"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(range(y_pred.shape[0]), y_pred,'r--')\n",
    "plt.plot(range(y_test.shape[0]), y_test,'g--')\n",
    "plt.legend(['preço previsto', 'preço real'])\n",
    "plt.ylabel('Preço')\n",
    "plt.xlabel('Índice')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "É possível observar no gráfico que o modelo, em alguns casos tende a prever preços mais baixos do que o preço real de venda."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Para saber o quanto o nosso modelo está explicando os dados, ou o quanto o modelo pode prever corretamente,\n",
    "iremos utilizar a medida r2, que é o coeficiente de determinação. Para isso iremos utilizar os dados da resposta\n",
    "na etapa de teste e os dados de resposta previstos."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2-score:  0.8231766135712899\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import r2_score\n",
    "\n",
    "print('R2-score: ', r2_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quanto mais próximo de 1, mais o modelo é capaz de prever corretamente.\n",
    "Colocando em porcentagem o resultado acima, podemos dizer que o nosso modelo\n",
    "prevê corretamente 82% dos preços."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Usando menos variáveis, a métrica r2 diminui"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R2-score:  0.775963917343572\n"
     ]
    }
   ],
   "source": [
    "new_x = df[['ano', 'preco_atual', 'kms_rodados', 'tipo_combustivel']]\n",
    "new_y = df[['preco_venda']]\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(new_x, new_y, test_size=0.2)\n",
    "\n",
    "model = LinearRegression()\n",
    "\n",
    "model.fit(x_train, y_train)\n",
    "\n",
    "new_y_pred = model.predict(x_test)\n",
    "\n",
    "print('R2-score: ', r2_score(y_test, new_y_pred))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}