session3_scientific_ecosystem/slides_session3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import xarray as xr\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# The Python Scientific Ecosystem\n",
    "\n",
    "<br><br><br><br><br><br><br><br><br><br><br>\n",
    "<div style=\"font-size:0.8em; color: #555\">2019-04-04 Session 3\n",
    "<span style=\"float:right\">Peter Regner</span>\n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# The Python Scientific Ecosystem\n",
    "\n",
    "<img alt=\"Python scientific ecosystem\" src=\"images/python-scientific-ecosystem.png\" style=\"width:800px;\">\n",
    "\n",
    "<small>Source: https://speakerdeck.com/jakevdp/the-state-of-the-stack-scipy-2015-keynote?slide=8</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# The Python Scientific Ecosystem\n",
    "\n",
    "- **numpy**: a `np.ndarray` represents a multidimensional, homogeneous array of fixed-size items (a Python list may contain objects of mixed types)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **scipy**: implementations of algorithms for data: filters, numerical optimization, interpolation, linalg, statistics, ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **pandas**: `pd.DataFrame` and `pd.Series`, great for things like time series"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **dask**: parallel computing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **xarray**: `xr.DataArray` is basically a `np.ndarray` with named dimensions with built-in support for `dask`, `xr.DataSet` similar to `pd.DataFrame`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Python vs. other languages and their applications\n",
    "\n",
    "<img src=\"images/language-comparison-datascience.png\" style=\"width:800px;\">\n",
    "\n",
    "<small>Source: https://speakerdeck.com/chdoig/the-state-of-python-for-data-science-pyss-2015?slide=22</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Runtime performance comparison\n",
    "\n",
    "\n",
    "<img src=\"images/benchmarks.svg\" style=\"width:800px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Note that this comparison is not very fair. The Python tests are mostly done with pure Python. One can easily speed-up tests by using numpy or other libraries. Adding ``@numba.jit`` to the Fibonacci test causes a speed up of factor 50 (tested with $n=30$).</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Source: https://julialang.org/benchmarks/</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# How to make code run fast?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<br><br>\n",
    "**TL;TR: Don't do loops!**\n",
    "\n",
    "(and use numpy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# How to make code run fast?\n",
    "\n",
    "- profile your code\n",
    "- use the best algorithm\n",
    "- use some fancy libraries\n",
    "- caching results\n",
    "- buy faster hardware\n",
    "- make things parallel (and buy more hardware)\n",
    "- use CUDA (and buy graphic cards)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Profile your code!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "<img src=\"images/optimization.png\" style=\"height:700px;\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Source: https://xkcd.com/1691/</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Profile your code!\n",
    "\n",
    "Profiling = measure runtime of a particular piece of code (often with the aim of optimization)\n",
    "\n",
    "The most simple approach to measure how long it takes to run a piece of code:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.30067896842956543"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import time\n",
    "\n",
    "def do_something_fancy():\n",
    "    time.sleep(0.3)\n",
    "\n",
    "t0 = time.time()\n",
    "\n",
    "do_something_fancy()\n",
    "\n",
    "elapsed_time = time.time() - t0\n",
    "\n",
    "elapsed_time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Profile your code!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8.91 ns ± 0.0469 ns per loop (mean ± std. dev. of 7 runs, 100000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit 39 + 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>For more options: https://docs.python.org/3/library/timeit.html</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Profile your code!\n",
    "\n",
    "Less popular, but really great to test parts of your code while running everything:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "from time import sleep\n",
    "import logging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0012243579985807"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from contexttimer import timer, Timer\n",
    "\n",
    "with Timer() as t:\n",
    "    sleep(1)\n",
    "    \n",
    "t.elapsed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "function sleep_one_sec execution time: 1.001 \n"
     ]
    }
   ],
   "source": [
    "logging.basicConfig(level=logging.INFO)\n",
    "\n",
    "@timer()\n",
    "def sleep_one_sec():\n",
    "    sleep(1)\n",
    "    \n",
    "sleep_one_sec()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Link: https://pypi.org/project/contexttimer/</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Profile your code: RunSankeRun\n",
    "\n",
    "\n",
    "<img src=\"images/run-snake-run.png\" style=\"600px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fast calculation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f(n):\n",
    "    if n  < 1:\n",
    "        return 0\n",
    "    elif n == 1:\n",
    "        return 1\n",
    "    \n",
    "    return f(n - 2) + f(n - 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[f(i) for i in range(10)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "This slide is just a reminder for how important good naming is ;-)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fast calculation of Fibonacci numbers\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fibonacci_number(n):\n",
    "    \"\"\"Return the n-th Fibonacci number (but you might need to be patiently waiting for the result).\"\"\"\n",
    "    if n  < 1:\n",
    "        return 0\n",
    "    elif n == 1:\n",
    "        return 1\n",
    "    \n",
    "    return fibonacci_number(n - 2) + fibonacci_number(n - 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[fibonacci_number(i) for i in range(10)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "The used algorithm with double recursion has exponential time complexity, but the problem could be solved in linear time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Asymptotic time complexity and the Big Oh\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Linear time complexity $O(n)$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- `x in l`  (for ``len(l) == n``)\n",
    "- looping over something of length n\n",
    "- max(`l`)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Logarithmic time complexity $O(log(n))$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "* binary search in a sorted list"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Asymptotic time complexity and the Big Oh\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Logarithmic time complexity $O(n \\cdot log(n))$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "* sorting a list"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Quadratic complexity $O(n^2)$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "* bubble sort (a slower sorting algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Asymptotic time complexity and the Big Oh\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Cubic time complexity $O(n^3)$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "* calculating the mean of wind speeds over `time x longitude resolution x latitude resolution`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Constant time $O(1)$:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "* `x[5]` access item in a list\n",
    "* `x in d` for `type(d) == dict`  (in average, not worst case)\n",
    "* `x in s` for `type(s) == set`  (in average, not worst case)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Source: https://wiki.python.org/moin/TimeComplexity</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fibonacci numbers: use good algorithms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fibonacci_number(n):\n",
    "    \"\"\"Return the n-th Fibonacci number (but you might need to be patiently waiting for the result).\"\"\"\n",
    "    if n  < 1:\n",
    "        return 0\n",
    "    elif n == 1:\n",
    "        return 1\n",
    "    \n",
    "    return fibonacci_number(n - 2) + fibonacci_number(n - 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[fibonacci_number(i) for i in range(10)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "The double recursive algorithm for calculating Fibonacci numbers is in $O(2^n)$, which is not tight.\n",
    "\n",
    "A the tight bound is $O(\\phi^n)$ where $\\phi = \\frac{1 + \\sqrt{5}}{2}$ is the golden ratio."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fibonacci numbers: use good algorithms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def fibonacci_number_linear(n):\n",
    "    \"\"\"Return the n-th Fibonacci number a bit faster.\"\"\"\n",
    "    element, subsequent = 0, 1\n",
    "    for i in range(n):\n",
    "        element, subsequent = subsequent, element + subsequent\n",
    "    \n",
    "    return subsequent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 1, 2, 3, 5, 8, 13, 21, 34, 55]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[fibonacci_number_linear(i) for i in range(10)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.1 s ± 7.56 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fibonacci_number(35)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.03 µs ± 17.7 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fibonacci_number_linear(35)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fast calculation of Fibonacci numbers: Caching"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import functools\n",
    "\n",
    "@functools.lru_cache()\n",
    "def fibonacci_number_memoized(n):\n",
    "    \"\"\"Return the n-th Fibonacci number (but you might need to be patiently waiting for the result).\"\"\"\n",
    "    if n  < 1:\n",
    "        return 0\n",
    "    elif n == 1:\n",
    "        return 1\n",
    "    \n",
    "    return fibonacci_number_memoized(n - 2) + fibonacci_number_memoized(n - 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[fibonacci_number_memoized(i) for i in range(10)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4.19 s ± 90.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fibonacci_number(35)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "74.2 ns ± 0.512 ns per loop (mean ± std. dev. of 7 runs, 10000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fibonacci_number_memoized(35)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Fast calculation of Fibonacci numbers: Caching\n",
    "\n",
    "\n",
    "I am currently looking for something like a file based `lru_cache()` to make it easy to play with different parameters for simulations that take ~10 minutes or so."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Speedup by use of some fancy libraries\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numba\n",
    "\n",
    "@numba.jit\n",
    "def fibonacci_number_numba(n):\n",
    "    \"\"\"Return the n-th Fibonacci number way faster.\"\"\"\n",
    "    element, subsequent = 0, 1\n",
    "    for i in range(n):\n",
    "        element, subsequent = subsequent, element + subsequent\n",
    "    \n",
    "    return subsequent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "270 ns ± 22.9 ns per loop (mean ± std. dev. of 7 runs, 1000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit fibonacci_number_numba(35)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Speedup by use of some fancy libraries\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "153 µs ± 4.98 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit sum(range(10_000))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Make it faster using some fancy library:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20.2 µs ± 479 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit np.sum(np.arange(10_000))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Make it faster using the Small Gauss, but kind of missing the purpose of the example... ;-)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9.31 ns ± 0.424 ns per loop (mean ± std. dev. of 7 runs, 100000000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit 10_000 * (10_000 - 1) // 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Speedup by other fancy alterantives to pure Python"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Alternative Python interpreters to the standard CPython implementation:\n",
    "\n",
    "- pypy: an Python interpreter written in Python\n",
    "- Cython: C-like performance, compiled"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Extensions in other languages: C, C++, Fortran, ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Other languages: Julia"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Buy faster hardware\n",
    "\n",
    "Okay, that one is easy, you need only enough money!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<img src=\"images/lotto.jpg\" style=\"width:800px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>\n",
    "Copyright: Österreichische Lotterien<br>\n",
    "Source: https://www.lotterien.at/presse/fotos/\n",
    "</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Buy faster hardware"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "source": [
    "<img src=\"images/trump-how_to_get_rich.jpg\" style=\"height:600px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Source: https://en.wikipedia.org/wiki/File:Trump_-_How_to_Get_Rich,_book_cover.jpg</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Make things parallel\n",
    "\n",
    "CPython is inherently single-core due to the global interpreter lock (GIL).\n",
    "\n",
    "Python threads actually slow down CPU bound tasks.\n",
    "\n",
    "<small>Source: https://callhub.io/understanding-python-gil/\n",
    "https://realpython.com/python-gil/\n",
    "https://wiki.python.org/moin/GlobalInterpreterLock</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Make things parallel\n",
    "\n",
    "- `multiprocessing`: part of the core library\n",
    "- `dask`: high comfort solution, but also low level support (MPI)\n",
    "- `joblib`, `torch.multiprocessing`: sorry, I don't know yet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Disclaimer: not sure if this list is complete."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Make things parallel: using big data\n",
    "\n",
    "Hadoop, Spark, ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<img src=\"images/map-reduce.png\" style=\"width:600px\">"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Make things run on the GPU\n",
    "\n",
    "- [numba](http://numba.pydata.org/)\n",
    "- [PyTorch](https://pytorch.org/)\n",
    "- [PyCUDA](https://mathema.tician.de/software/pycuda/)\n",
    "\n",
    "For machine learning:\n",
    "- Theano\n",
    "- TensorFlow\n",
    "- ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Latency Numbers Every Programmer Should Know  (~2012)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small><pre>\n",
    "    L1 cache reference                           0.5 ns\n",
    "    Branch mispredict                            5   ns\n",
    "    L2 cache reference                           7   ns                      14x L1 cache\n",
    "    Mutex lock/unlock                           25   ns\n",
    "    Main memory reference                      100   ns                      20x L2 cache, 200x L1 cache\n",
    "    Compress 1K bytes with Zippy             3,000   ns        3 us\n",
    "    Send 1K bytes over 1 Gbps network       10,000   ns       10 us\n",
    "    Read 4K randomly from SSD*             150,000   ns      150 us          ~1GB/sec SSD\n",
    "    Read 1 MB sequentially from memory     250,000   ns      250 us\n",
    "    Round trip within same datacenter      500,000   ns      500 us\n",
    "    Read 1 MB sequentially from SSD*     1,000,000   ns    1,000 us    1 ms  ~1GB/sec SSD, 4X memory\n",
    "    Disk seek                           10,000,000   ns   10,000 us   10 ms  20x datacenter roundtrip\n",
    "    Read 1 MB sequentially from disk    20,000,000   ns   20,000 us   20 ms  80x memory, 20X SSD\n",
    "    Send packet CA->Netherlands->CA    150,000,000   ns  150,000 us  150 ms\n",
    "\n",
    "    Notes\n",
    "    -----\n",
    "    1 ns = 10^-9 seconds\n",
    "    1 us = 10^-6 seconds = 1,000 ns\n",
    "    1 ms = 10^-3 seconds = 1,000 us = 1,000,000 ns\n",
    "<pre></small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Reading 1MB from memory vs from disk (with seek):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "120.0"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(20e-3 + 10e-3) / 250e-6"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Source: \n",
    " - https://gist.github.com/jboner/2841832\n",
    " - https://gist.github.com/GLMeece/b00c9c97a06a957af7426b1be5bc8be6\n",
    " - http://norvig.com/21-days.html#answers\n",
    " - https://people.eecs.berkeley.edu/~rcs/research/interactive_latency.html"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Don't forget about numerics!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "0.1 + 0.2 == 0.3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Questions?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Exercise\n",
    "\n",
    "Try to find out why slow_average.py is slow and make it fast!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Tips and tricks\n",
    "\n",
    "Python can be used:\n",
    "\n",
    "- as Python **program/script/module**, i.e. in `*.py` files: long running calculations, reproducible results, avoiding code duplication"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **IPython** terminal: try out something quickly (and throw it away)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **Jupyter notebook** and Jupyter lab: make drafts and summarize exploration of data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- Scientific **cell mode**: PyCharm or Spyder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- **unit tests**: instead of testing manually, run your your tests"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "All of them are also available directly in many IDEs/editors!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Tips and tricks\n",
    "\n",
    "IPython:\n",
    "\n",
    "Use `CTRL` + `R` to find older commands\n",
    "\n",
    "...or use more fancy commands:\n",
    "```\n",
    ">>> %history ~1/1-100\n",
    ">>> %edit ~1/1-100\n",
    ">>> %notebook /tmp/my_new_jupyter_notebook.ipynb\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Getting help:\n",
    "    \n",
    "```\n",
    ">>> np.mean?\n",
    "```\n",
    "\n",
    "Display source code:\n",
    "```\n",
    ">>> np.mean??\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Tips and tricks\n",
    "\n",
    "Jupyter notebooks: see [example-notebook.ipynb](example-notebook.ipynb)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Tips and tricks\n",
    "\n",
    "Editor must have features:\n",
    "\n",
    "- jump to definition\n",
    "- syntax highlighting\n",
    "- linting\n",
    "\n",
    "Really helpful shortcuts:\n",
    "- search for word under cursor \\*\n",
    "- duplicate line\n",
    "- empty line/above below\n",
    "- block select\n",
    "- macros\n",
    "\n",
    "Also helpful:\n",
    "- shortcuts to open \n",
    "- tools for search in all files in a folder, in linux `grep` (or faster: `git grep`, `ripgrep`, `ag`, ...)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Tips and tricks\n",
    "\n",
    "Keep an eye on memory usage and CPU when running code on your machine!\n",
    "\n",
    "For windows, maybe this helps: https://entropy6.com/xmeters/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Numpy\n",
    "\n",
    "...introduces the type `np.ndarray`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 6, 7])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = np.arange(8)\n",
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8,)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 1, 2, 3],\n",
       "       [4, 5, 6, 7]])"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a.reshape(2, 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a.reshape(2,4).ndim"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Numpy\n",
    "\n",
    "Functions to create an `np.ndarray`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1, 2],\n",
       "       [3, 4]])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# note that we use np.array:\n",
    "np.array([[1, 2],\n",
    "          [3, 4]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1., 1.],\n",
       "       [1., 1.]])"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.ones((2,2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Numpy\n",
    "\n",
    "Functions to create an `np.ndarray`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[False, False],\n",
       "       [False, False]])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.zeros((2,2), dtype=np.bool)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.   , 0.   , 0.   , 1.875], dtype=float32)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.empty(4, dtype=np.float32)  # amazingly fast but dangerously uninitialized!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "`np.diag`, `np.tri`, ..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: Broadcasting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5., 5., 5.],\n",
       "       [5., 5., 5.],\n",
       "       [5., 5., 5.]])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "5 * np.ones((3,3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.]],\n",
       "\n",
       "       [[11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.]],\n",
       "\n",
       "       [[11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.],\n",
       "        [11., 22.]]])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.zeros((3, 4, 2)) + np.array([11, 22])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: Broadcasting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "```\n",
    "A      (3d array):  15 x 3 x 5\n",
    "B      (2d array):       3 x 5\n",
    "Result (3d array):  15 x 3 x 5\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4, 1)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "factor = np.array([[11], [22], [33], [44]])\n",
    "factor.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 4, 2)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result = np.ones((3, 4, 2)) * factor\n",
    "result.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[11., 11.],\n",
       "        [22., 22.],\n",
       "        [33., 33.],\n",
       "        [44., 44.]],\n",
       "\n",
       "       [[11., 11.],\n",
       "        [22., 22.],\n",
       "        [33., 33.],\n",
       "        [44., 44.]],\n",
       "\n",
       "       [[11., 11.],\n",
       "        [22., 22.],\n",
       "        [33., 33.],\n",
       "        [44., 44.]]])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<small>Link: https://docs.scipy.org/doc/numpy/user/basics.broadcasting.html</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Broadcasting: Numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "latitude_resolution = 100\n",
    "longitude_resolution = 110\n",
    "time_resolution = 365\n",
    "\n",
    "# in m/s, postfixing with _ms does not work well, because ms is miliseconds... :-/\n",
    "wind_speeds = np.random.weibull(10, size=(latitude_resolution, longitude_resolution, time_resolution))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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.hist(wind_speeds.flatten(), density=True, bins=50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Broadcasting: Numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f91fd597c50>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "# m/s --> kW\n",
    "wind_to_power = np.random.normal(300, scale=20, size=(latitude_resolution, longitude_resolution))\n",
    "plt.imshow(wind_to_power)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Broadcasting: Numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "operands could not be broadcast together with shapes (100,110) (100,110,365) ",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-44-caa40600a1f1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mwind_to_power\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mwind_speeds\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m: operands could not be broadcast together with shapes (100,110) (100,110,365) "
     ]
    }
   ],
   "source": [
    "wind_to_power * wind_speeds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 110, 365)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_speeds.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 110)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_to_power.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Broadcasting: Numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "wind_speeds = np.rollaxis(wind_speeds, 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(365, 100, 110)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_speeds.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 110)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_to_power.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "average_power = (wind_to_power * wind_speeds).mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Xarray has named dimensions!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "wind_speeds_raw = np.random.weibull(10, size=(latitude_resolution, longitude_resolution, time_resolution))\n",
    "wind_speeds = xr.DataArray(wind_speeds_raw, dims=('latitude', 'longitude', 'time'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<xarray.DataArray (latitude: 100, longitude: 110, time: 365)>\n",
       "array([[[0.894741, 0.990923, ..., 0.899981, 1.02659 ],\n",
       "        [1.020849, 0.624725, ..., 1.005816, 0.784432],\n",
       "        ...,\n",
       "        [0.811398, 0.986918, ..., 0.939516, 1.110318],\n",
       "        [0.976799, 0.977067, ..., 1.012583, 0.917623]],\n",
       "\n",
       "       [[0.993984, 1.035443, ..., 1.069029, 0.649291],\n",
       "        [0.689977, 0.811261, ..., 0.951617, 1.186624],\n",
       "        ...,\n",
       "        [0.779403, 0.78594 , ..., 1.111433, 0.977469],\n",
       "        [0.801538, 0.939266, ..., 0.780707, 0.872474]],\n",
       "\n",
       "       ...,\n",
       "\n",
       "       [[0.886154, 0.922524, ..., 1.112934, 1.119962],\n",
       "        [1.012106, 0.846248, ..., 1.053159, 1.054678],\n",
       "        ...,\n",
       "        [0.85447 , 0.881155, ..., 0.844021, 1.054484],\n",
       "        [0.831682, 1.071129, ..., 0.94867 , 1.095568]],\n",
       "\n",
       "       [[0.929042, 0.984445, ..., 1.034028, 0.978367],\n",
       "        [1.021332, 1.025009, ..., 0.888362, 0.806673],\n",
       "        ...,\n",
       "        [1.122647, 1.001822, ..., 0.902706, 0.943693],\n",
       "        [0.991303, 0.947753, ..., 0.88277 , 0.911381]]])\n",
       "Dimensions without coordinates: latitude, longitude, time"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_speeds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "wind_to_power_raw = np.random.normal(300, scale=20, size=(latitude_resolution, longitude_resolution))\n",
    "wind_to_power = xr.DataArray(wind_to_power_raw, dims=('latitude', 'longitude'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<xarray.DataArray (latitude: 100, longitude: 110)>\n",
       "array([[267.155648, 260.649821, 304.058154, ..., 298.093269, 262.57353 ,\n",
       "        278.500393],\n",
       "       [300.75415 , 258.107828, 291.038777, ..., 288.448435, 242.067571,\n",
       "        310.542559],\n",
       "       [290.045851, 285.304637, 290.781913, ..., 291.486622, 315.935329,\n",
       "        259.97107 ],\n",
       "       ...,\n",
       "       [291.441859, 314.460253, 278.068775, ..., 299.31902 , 286.979371,\n",
       "        278.728631],\n",
       "       [293.0093  , 309.4248  , 261.796419, ..., 272.178113, 279.230288,\n",
       "        280.699875],\n",
       "       [306.753309, 322.732434, 299.182011, ..., 278.764251, 277.416258,\n",
       "        278.512955]])\n",
       "Dimensions without coordinates: latitude, longitude"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(wind_to_power * wind_speeds).mean(dim='time')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Numpy: this is how you make things fast!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "a = 39 * np.ones(10**5)\n",
    "b = 2 * np.ones(10**5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30.7 ms ± 952 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit [i + j for i, j in zip(a, b)]  # list comprehension is still better than a loop!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50.1 µs ± 4.53 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit a + b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: Slicing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d = np.arange(5)\n",
    "d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 3])"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d[2:4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 2, 4])"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d[::2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4, 3, 2, 1, 0])"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "d[::-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1., 1., 1., 1.],\n",
       "       [1., 1., 1., 1.],\n",
       "       [1., 1., 1., 1.]])"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_ones = np.ones((3,5))\n",
    "some_ones[:, :-1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: when is data copied?\n",
    "\n",
    "It's complicated but really important... :-/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "some_numbers = np.arange(5)\n",
    "less_numbers = some_numbers[:-2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_numbers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2])"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "less_numbers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "140265030355984"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "id(some_numbers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "140265030356464"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "id(less_numbers)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: when is data copied?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "some_numbers[0] = 42"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  2])"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "less_numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Slicing returns a view!**\n",
    "\n",
    "(something like two names pointing to the same object, but now there are two objects pointing to the same memory)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "slideshow": {
     "slide_type": "skip"
    }
   },
   "outputs": [],
   "source": [
    "class Shrug:\n",
    "    def __repr__(self):\n",
    "        return '¯\\_(ツ)_/¯'\n",
    "    \n",
    "(ツ) = Shrug()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "¯\\_(ツ)_/¯"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(ツ)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<small>Source: https://scipy-cookbook.readthedocs.io/items/ViewsVsCopies.html</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: when is data copied?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  2,  3,  4])"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "Fancy indexing returns a copy, not a view:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  2,  4])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fancy_indexing = some_numbers[[0, 2, -1]]\n",
    "fancy_indexing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([84,  2,  4])"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fancy_indexing[0] *= 2\n",
    "fancy_indexing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  2,  3,  4])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "#  Numpy: when is data copied?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  2,  3,  4])"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "source": [
    "Same here:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  3])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fancy_indexing = some_numbers[[True, True, False, True, False]]\n",
    "fancy_indexing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([84,  1,  3])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fancy_indexing[0] *= 2\n",
    "fancy_indexing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([42,  1,  2,  3,  4])"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_numbers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<small>More: https://jakevdp.github.io/PythonDataScienceHandbook/02.07-fancy-indexing.html</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Plotting\n",
    "\n",
    "- not so much interactive\n",
    "    - matplotlib\n",
    "    - seaborn\n",
    "\n",
    "- interactive\n",
    "    - plotly\n",
    "    - bokeh"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Matplotlib, Pyplot, Pylab"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- pyplot is an easier way to use matplotlib, but it can be dangerous (if used in parallel)\n",
    "- for datascience pyplot should be good enough for most data-science or science-with-data applications\n",
    "- pylab import pyblot and numpy for you, but it does a `from numpy import *` which overwrites e.g. the Python `sum()`"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "<small>Source: https://matplotlib.org/faq/usage_faq.html#matplotlib-pyplot-and-pylab-how-are-they-related</small>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Nice examples with pandas, matplotlib and seaborn\n",
    "\n",
    "- https://github.com/caichinger/slides/blob/master/2019-03-19__DjangoVienna/1_groupby.ipynb\n",
    "- https://github.com/caichinger/slides/blob/master/2019-03-19__DjangoVienna/2_timeseries.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Exercise: plot cos(x) * sin(y)\n",
    "\n",
    "Plot  $f(x,y) := cos(x) * sin(y)$!\n",
    "\n",
    "(for $x,y \\in \\mathbb{R}, -10 \\leq x \\leq 10, -10 \\leq y \\leq 10$ with a resolution of $300 \\times 300$)\n",
    "\n",
    "Use the functions:\n",
    "- `np.linspace`\n",
    "- `np.meshgrid`\n",
    "- `np.cos`, `np.sin`\n",
    "- `plt.imshow`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "resolution = 300\n",
    "coords = np.linspace(-10, 10, num=resolution)\n",
    "xx, yy = np.meshgrid(coords, coords)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Exercise: plot cos(x) * cos(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "slideshow": {
     "slide_type": ""
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-10.        ,  -9.93311037,  -9.86622074,  -9.7993311 ],\n",
       "       [-10.        ,  -9.93311037,  -9.86622074,  -9.7993311 ],\n",
       "       [-10.        ,  -9.93311037,  -9.86622074,  -9.7993311 ],\n",
       "       [-10.        ,  -9.93311037,  -9.86622074,  -9.7993311 ]])"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xx[:4, :4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-10.        , -10.        , -10.        , -10.        ],\n",
       "       [ -9.93311037,  -9.93311037,  -9.93311037,  -9.93311037],\n",
       "       [ -9.86622074,  -9.86622074,  -9.86622074,  -9.86622074],\n",
       "       [ -9.7993311 ,  -9.7993311 ,  -9.7993311 ,  -9.7993311 ]])"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "yy[:4, :4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f91fd4ccef0>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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.imshow(np.cos(xx) * np.sin(yy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Exercise\n",
    "\n",
    "Imagine an API which allows you to query locations of some object of interest (wind turbines, gas stations, ...). For some given location $x \\in \\mathbb{R}$ (=API input), the API will tell you the closest `num_neighbors` $\\in \\mathbb{N}$ object locations."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "We are interested in getting all locations. For which/how many locations do we need to query?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "source": [
    "- $N$ `locations` in 2D (picked randomly (uniform) from $[0,1] \\times [0,1] \\subseteq \\mathbb{R}^2$\n",
    "- `num_neighbors` is the number of locations an API request will return\n",
    "- plot the map of `locations` with a resolution of `map_resolution` $\\times$ `map_resolution` and as background mark positions with equal API answers with the same color"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- ``np.linspace(start, stop, num=NUM)``\n",
    "- ``xx, yy = np.meshgrid(x, y)``\n",
    "- ``np.linalg.norm(x, axis=0)``"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Exercise: solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "N = 20\n",
    "margin = 0.3\n",
    "map_resolution = 900\n",
    "num_neighbours = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "np.random.seed(42)   # deterministic randomess, yeah!\n",
    "locations = np.random.rand(2, N)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "coords = np.linspace(-margin, margin + 1, num=map_resolution)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "xx, yy = np.meshgrid(coords, coords)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [],
   "source": [
    "# axis=0 <-- canvas x\n",
    "# axis=1 <-- canvas y\n",
    "# axis=2 <-- locations\n",
    "distances = np.linalg.norm(locations.T - np.dstack((xx,yy))[:, :, np.newaxis, :], axis=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "canvas = np.argsort(distances, axis=2)[:, :, :num_neighbours]\n",
    "canvas = np.sort(canvas, axis=2)\n",
    "canvas = np.sum(N**np.arange(num_neighbours)[::-1] * canvas, axis=2) \n",
    "canvas = canvas / canvas.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x1152 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(canvas, extent=(-margin, 1 + margin, 1 + margin, -margin), cmap='gist_ncar')\n",
    "ax = plt.plot(locations[0], locations[1], 'or')\n",
    "\n",
    "ax[0].figure.set_figheight(16)\n",
    "ax[0].figure.set_figwidth(12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "The same thing in R:\n",
    "\n",
    "```\n",
    "set.seed(42)\n",
    "\n",
    "canvas_m = 0.3\n",
    "canvas_r = 900\n",
    "\n",
    "locations_n = 25\n",
    "neighbours_n = 4\n",
    "\n",
    "locations = matrix(runif(2*locations_n), ncol=2, \n",
    "                  dimnames=list(NULL, c('x', 'y')))\n",
    "\n",
    "canvas_l = seq(-canvas_m, canvas_m + 1, length=canvas_r)\n",
    "canvas_c = as.matrix(expand.grid(x = canvas_l, y = canvas_l))\n",
    "\n",
    "canvas_c = cbind(canvas_c, v = apply(canvas_c, 1, function(p) {\n",
    "    d = sqrt((locations[, 1] - p[1])**2 + (locations[, 2] - p[2])**2)\n",
    "    sum(locations_n**order(d)[1:neighbours_n])\n",
    "}))\n",
    "\n",
    "library(ggplot2)\n",
    "ggplot(as.data.frame(canvas_c), aes(x = x, y = y)) +\n",
    "    geom_tile(aes(fill = factor(v))) + \n",
    "    geom_point(data = as.data.frame(locations), color = 'red') +\n",
    "    scale_fill_discrete(guide=FALSE)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Not covered in this workshop\n",
    "\n",
    "- f-strings\n",
    "- decorators\n",
    "- context managers\n",
    "- iterators\n",
    "- generators\n",
    "- data scraping\n",
    "- packaging and package managers\n",
    "- virtualenv, conda env, pipenv, ...\n",
    "- `from __future__ import braces`"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "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": "3.6.7"
  },
  "rise": {
   "center": false,
   "controls": false,
   "scroll": true,
   "slideNumber": "c/t",
   "theme": "serif",
   "transition": "none"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}