{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "4d2a8b6c",
   "metadata": {},
   "source": [
    "#### Database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7be9eeff",
   "metadata": {},
   "outputs": [],
   "source": [
    "PROJECT_PATH = '/home/md/Work/ligalytics/leagues_stable/'\n",
    "import os, sys\n",
    "sys.path.insert(0, PROJECT_PATH)\n",
    "os.environ.setdefault(\"DJANGO_SETTINGS_MODULE\", \"leagues.settings\")\n",
    "\n",
    "from leagues import settings\n",
    "settings.DATABASES['default']['NAME'] = PROJECT_PATH+'/db.sqlite3'\n",
    "\n",
    "import django\n",
    "django.setup()\n",
    "\n",
    "from scheduler.models import *\n",
    "from common.functions import distanceInKmByGPS\n",
    "season = Season.objects.filter(nicename=\"Imported: Benchmark Season\").first()\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from django.db.models import Count, F, Value\n",
    "games = Game.objects.filter(season=season)\n",
    "df = pd.DataFrame.from_records(games.values())\n",
    "games = Game.objects.filter(season=season).annotate(\n",
    "    home=F('homeTeam__shortname'),\n",
    "    away=F('awayTeam__shortname'),\n",
    "    home_lat=F('homeTeam__latitude'),\n",
    "    home_lon=F('homeTeam__longitude'),\n",
    "    home_attr=F('homeTeam__attractivity'),\n",
    "    away_lat=F('awayTeam__latitude'),\n",
    "    away_lon=F('awayTeam__longitude'),\n",
    "    away_attr=F('awayTeam__attractivity')\n",
    ").values()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc191792",
   "metadata": {},
   "source": [
    "#### Dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1e404cf8",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "# create dataset\n",
    "df = pd.DataFrame.from_records(games.values())\n",
    "\n",
    "# pivots\n",
    "pivot_homeTeam_mean = df.pivot_table('attendance','homeTeam_id',aggfunc='mean')\n",
    "pivot_homeTeam_max = df.pivot_table('attendance','homeTeam_id',aggfunc='max')\n",
    "\n",
    "# add more features\n",
    "df['weekday'] = df.apply(lambda r: r['date'].weekday(), axis=1)\n",
    "df['day'] = df.apply(lambda r: r['date'].day, axis=1)\n",
    "df['month'] = df.apply(lambda r: r['date'].month, axis=1)\n",
    "df['year'] = df.apply(lambda r: r['date'].year, axis=1)\n",
    "df['distance'] = df.apply(lambda r: distanceInKmByGPS(r['home_lon'],r['home_lat'],r['away_lon'],r['away_lat']), axis=1)\n",
    "df['weekend'] = df.apply(lambda r: int(r['weekday'] in [6,7]), axis=1)\n",
    "df['winter_season'] = df.apply(lambda r: int(r['month'] in [1,2,3,10,11,12]), axis=1)\n",
    "df['home_base'] = df.apply(lambda r: pivot_homeTeam_mean.loc[r['homeTeam_id'],'attendance'], axis=1)\n",
    "df['stadium_size'] = df.apply(lambda r: pivot_homeTeam_max.loc[r['homeTeam_id'],'attendance'], axis=1)\n",
    "\n",
    "# one hot encoding\n",
    "ohe_fields = ['time', 'historic_season']\n",
    "\n",
    "for field in ohe_fields:\n",
    "    ohe = OneHotEncoder()\n",
    "    transformed = ohe.fit_transform(df[[field]])\n",
    "    df[ohe.categories_[0]] = transformed.toarray()\n",
    "\n",
    "# sort label to last index\n",
    "cols = list(df.columns)\n",
    "cols.append(cols.pop(cols.index('attendance')))\n",
    "df = df[cols]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e69d24dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Importing Libraries\n",
    "import numpy as np # linear algebra\n",
    "import pandas as pd # data processing\n",
    "import matplotlib.pyplot as plt # plotting library\n",
    "from sklearn.model_selection import train_test_split,cross_val_score, cross_val_predict\n",
    "from sklearn import metrics\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.preprocessing import PolynomialFeatures\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e2ea08e5",
   "metadata": {},
   "source": [
    "#### Train/Test Data - Normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "74e12f87",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "\n",
    "remove_columns = ['season_id', 'resultEntered', 'reversible', 'reschedule', 'homeGoals', 'awayGoals',\n",
    "                  'homeGoals2', 'awayGoals2', 'homeGoals3', 'awayGoals3', 'home', 'away', 'date', 'time', 'historic_season', 'id', 'homeTeam_id', 'awayTeam_id']\n",
    "feature_cols = list(set(df.columns[:-1]) - set(remove_columns))\n",
    "# feature_cols = ['weekday','weekend','home_base','distance','winter_season']\n",
    "label = 'attendance'\n",
    "\n",
    "\n",
    "X = df[feature_cols]  # Features\n",
    "y = df[label]  # Target variable\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.3, random_state=1)  # 70% training and 30% test"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94ade4b4",
   "metadata": {},
   "source": [
    "#### Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "4c9bdd0d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "FITTING...done\n",
      "VISUALIZE\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pydotplus\n",
    "from six import StringIO\n",
    "from sklearn.tree import export_graphviz\n",
    "from sklearn.tree import DecisionTreeRegressor \n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "# Create Decision Tree classifer object\n",
    "regr = DecisionTreeRegressor(max_depth=5, random_state=1234)\n",
    "\n",
    "# Train Decision Tree Classifer\n",
    "print(\"FITTING...\", end=\"\")\n",
    "regr = regr.fit(X_train, y_train)\n",
    "print(\"done\")\n",
    "\n",
    "# Predict the response for test dataset\n",
    "y_pred = regr.predict(X_test)\n",
    "\n",
    "print(\"VISUALIZE\")\n",
    "dot_data = StringIO()\n",
    "export_graphviz(regr, out_file=dot_data,\n",
    "                filled=True, rounded=True,\n",
    "                special_characters=True, feature_names=feature_cols)\n",
    "graph = pydotplus.graph_from_dot_data(dot_data.getvalue())\n",
    "graph.write_png('attendance.png')\n",
    "# Image(graph.create_png())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7.13 ('leagues')",
   "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": "3.7.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "a07b7f3079ca8c056705d3c757c4f3f92f9509f33eeab9ad5420dacec37bc01a"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}