notebooks/2-basic-usage.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Basic Usage\n",
    "This notebook illustrates how the default pipeline can be used to compute xG values. This default pipeline is suitable for computing xG, but by no means the best possible model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import itertools\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "from soccer_xg.api import DataApi\n",
    "import soccer_xg.xg as xg\n",
    "from soccer_xg.utils import sample_temporal"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dataset\n",
    "dir_data = \"../data\"\n",
    "provider = 'wyscout_opensource'\n",
    "leagues = ['ENG', 'ESP', 'ITA', 'GER', 'FRA']\n",
    "seasons = ['1718']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load the data\n",
    "Make sure to run the `1-load-and-convert-wyscout-data` notebook first to download the data set and convert it to the SPADL format."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "api = DataApi([f\"{dir_data}/{provider}/spadl-{provider}-{l}-{s}.h5\" \n",
    "        for (l,s) in itertools.product(leagues, seasons)])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model learning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, create an instance of the default xG model. This is basically a wrapper around three seperate Scikit-learn pipelines (a seperate pipeline for open play shots, freekicks and penalties) for computing xG values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = xg.XGModel()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once the modeling pipeline is set up, historical data needs to be fed into it in order to\n",
    "\"fit\" the model so that it can then be used to predict future results.\n",
    "This `train` method implements a simple wrapper around the core Scikit-learn functionality\n",
    "which does this. We only use the data from ESP, ITA, FRA and GER to train the model. ENG is reserved for validating the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generating features: 100%|██████████| 1446/1446 [06:26<00:00,  3.75it/s]\n",
      "Generating labels: 100%|██████████| 1446/1446 [00:33<00:00, 43.29it/s]\n",
      "Generating features: 100%|██████████| 1446/1446 [04:27<00:00,  5.41it/s]\n",
      "Generating labels: 100%|██████████| 1446/1446 [00:33<00:00, 43.78it/s]\n"
     ]
    }
   ],
   "source": [
    "model.train(api, training_seasons=[('ESP', '1718'), ('ITA', '1718'), ('FRA', '1718'), ('GER', '1718')])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once a modeling pipeline is trained, a different dataset must be fed into the trained model\n",
    "to validate the quality of the fit. Ideally, this validation set would consist out of more recent data from the same leagues as used for training. However, since the wyscout dataset contains only one season of data, we use the data from ENG to validate the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generating labels: 100%|██████████| 380/380 [00:11<00:00, 33.62it/s]\n",
      "Generating features: 100%|██████████| 380/380 [02:01<00:00,  3.13it/s]\n",
      "Generating features: 100%|██████████| 380/380 [01:16<00:00,  4.95it/s]\n",
      "Generating features: 100%|██████████| 380/380 [00:11<00:00, 33.75it/s]\n",
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'max_dev': 30.926317206053028,\n",
       " 'residual_area': 823.2684757471029,\n",
       " 'roc': 0.7871783922219415,\n",
       " 'brier': 0.08585818466560044,\n",
       " 'ece': 0.010637320991901903,\n",
       " 'ace': 0.014368773223275452,\n",
       " 'fig': <Figure size 720x360 with 3 Axes>}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.validate(api, validation_seasons=[('ENG', '1718')])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Computing xG values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Once trained and validated, the model can be used to compute xG values. For example, the cell below computes the xG values of all shots in the 17/18 Premier League season."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generating features: 100%|██████████| 380/380 [02:00<00:00,  3.15it/s]\n",
      "Generating features: 100%|██████████| 380/380 [01:17<00:00,  4.91it/s]\n",
      "Generating features: 100%|██████████| 380/380 [00:10<00:00, 35.75it/s]\n"
     ]
    }
   ],
   "source": [
    "games_eng_1718 = api.games[api.games.competition_id == 'ENG']\n",
    "xg_eng_1718 = model.estimate(api, game_ids=games_eng_1718.index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These can be aggregated to compute the total xG value of each player in the 17/18 season."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "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>xG</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>short_name</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>H. Kane</th>\n",
       "      <td>28.714433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mohamed Salah</th>\n",
       "      <td>22.768716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>S. Agüero</th>\n",
       "      <td>18.095464</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>J. Vardy</th>\n",
       "      <td>17.587558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>R. Lukaku</th>\n",
       "      <td>17.147181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>R. Sterling</th>\n",
       "      <td>15.968029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A. Lacazette</th>\n",
       "      <td>14.283950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>G. Murray</th>\n",
       "      <td>14.157180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gabriel Jesus</th>\n",
       "      <td>13.436479</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>L. Milivojević</th>\n",
       "      <td>11.952281</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       xG\n",
       "short_name               \n",
       "H. Kane         28.714433\n",
       "Mohamed Salah   22.768716\n",
       "S. Agüero       18.095464\n",
       "J. Vardy        17.587558\n",
       "R. Lukaku       17.147181\n",
       "R. Sterling     15.968029\n",
       "A. Lacazette    14.283950\n",
       "G. Murray       14.157180\n",
       "Gabriel Jesus   13.436479\n",
       "L. Milivojević  11.952281"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(xg_eng_1718\n",
    " # find player id for each shot\n",
    " .join(api.actions.set_index(['game_id', 'action_id']), how=\"left\")\n",
    " # sum xG values for each player\n",
    " .groupby('player_id')[['xG']].sum()\n",
    " # find player name\n",
    " .join(api.players[['short_name']], how=\"left\")\n",
    " # get top 10\n",
    " .sort_values(by='xG', ascending=False).head(n=10)\n",
    " # clean up the list\n",
    " .set_index('short_name', drop=True)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "soccer_dataprovider_comparison",
   "language": "python",
   "name": "soccer_dataprovider_comparison"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}