icehockey_matchprediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict ice hockey game result with machine learning\n",
    "\n",
    "The use of data analytics and statistics to predict outcome of sport events is nothing new and is of course used by the betting companies. You look at trends, table position, head-to-head results etc. And for the individual players, dont forget the most important aspect: This vague \"feeling\".  \n",
    "\n",
    "My \"feeling\" is normally not very accurate, and thus I wanted to set up a machine learning model that predicted the outcome of ice hockey games. Maybe this could help me to improve accuracy in my predictions, and not loose that much money. \n",
    "\n",
    "This notebook describes how I developed a random forest classifier to predict outcome of ice hockey games in the swedish 1st and 2nd league. I will also apply the model to the season results to see how it would have performed.\n",
    "\n",
    "I will use two datasets collected with the package ```swehockey_scraper``` which collects data from www.stats.swehockey.se. \n",
    "Details about the package is found here [github project for ice-hockey-analysis](https://github.com/msjoelin/swehockey_scraper)\n",
    "\n",
    "Steps:\n",
    "1. Import data\n",
    "2. Exploratory analysis\n",
    "3. Prepare data for the machine learning model\n",
    "4. Train model\n",
    "5. Evaluate performance\n",
    "6. Apply model to result current season and \n",
    "\n",
    "So let's start!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1) Import  packages and data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "# sklearn will be used to build the model\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split \n",
    "from sklearn import preprocessing\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import classification_report\n",
    "\n",
    "# To save the model\n",
    "import pickle\n",
    "\n",
    "# For visualization\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9184  matches imported\n"
     ]
    },
    {
     "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>date</th>\n",
       "      <th>game</th>\n",
       "      <th>score</th>\n",
       "      <th>league</th>\n",
       "      <th>season</th>\n",
       "      <th>game_id</th>\n",
       "      <th>home</th>\n",
       "      <th>away</th>\n",
       "      <th>score_home</th>\n",
       "      <th>score_away</th>\n",
       "      <th>result</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2019-09-14</td>\n",
       "      <td>Frölunda HC  - Brynäs IF</td>\n",
       "      <td>2 - 5</td>\n",
       "      <td>shl</td>\n",
       "      <td>2019/20</td>\n",
       "      <td>441805</td>\n",
       "      <td>Frölunda HC</td>\n",
       "      <td>Brynäs IF</td>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>away</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2019-09-14</td>\n",
       "      <td>Skellefteå AIK  - IK Oskarshamn</td>\n",
       "      <td>5 - 4</td>\n",
       "      <td>shl</td>\n",
       "      <td>2019/20</td>\n",
       "      <td>441808</td>\n",
       "      <td>Skellefteå AIK</td>\n",
       "      <td>IK Oskarshamn</td>\n",
       "      <td>5</td>\n",
       "      <td>4</td>\n",
       "      <td>draw</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2019-09-14</td>\n",
       "      <td>Leksands IF  - IF Malmö Redhawks</td>\n",
       "      <td>5 - 2</td>\n",
       "      <td>shl</td>\n",
       "      <td>2019/20</td>\n",
       "      <td>441809</td>\n",
       "      <td>Leksands IF</td>\n",
       "      <td>IF Malmö Redhawks</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>home</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2019-09-14</td>\n",
       "      <td>Djurgårdens IF  - Linköping HC</td>\n",
       "      <td>4 - 2</td>\n",
       "      <td>shl</td>\n",
       "      <td>2019/20</td>\n",
       "      <td>441803</td>\n",
       "      <td>Djurgårdens IF</td>\n",
       "      <td>Linköping HC</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>home</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2019-09-14</td>\n",
       "      <td>Färjestad BK  - Örebro HK</td>\n",
       "      <td>4 - 6</td>\n",
       "      <td>shl</td>\n",
       "      <td>2019/20</td>\n",
       "      <td>441804</td>\n",
       "      <td>Färjestad BK</td>\n",
       "      <td>Örebro HK</td>\n",
       "      <td>4</td>\n",
       "      <td>6</td>\n",
       "      <td>away</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date                              game  score league   season  \\\n",
       "0  2019-09-14          Frölunda HC  - Brynäs IF  2 - 5    shl  2019/20   \n",
       "1  2019-09-14   Skellefteå AIK  - IK Oskarshamn  5 - 4    shl  2019/20   \n",
       "2  2019-09-14  Leksands IF  - IF Malmö Redhawks  5 - 2    shl  2019/20   \n",
       "3  2019-09-14    Djurgårdens IF  - Linköping HC  4 - 2    shl  2019/20   \n",
       "4  2019-09-14         Färjestad BK  - Örebro HK  4 - 6    shl  2019/20   \n",
       "\n",
       "   game_id            home               away  score_home  score_away result  \n",
       "0   441805     Frölunda HC          Brynäs IF           2           5   away  \n",
       "1   441808  Skellefteå AIK      IK Oskarshamn           5           4   draw  \n",
       "2   441809     Leksands IF  IF Malmö Redhawks           5           2   home  \n",
       "3   441803  Djurgårdens IF       Linköping HC           4           2   home  \n",
       "4   441804    Färjestad BK          Örebro HK           4           6   away  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Import game data. Only keep relevant columns\n",
    "url_games = 'https://raw.githubusercontent.com/msjoelin/icehockey_analysis/master/data/df_games.csv'\n",
    "\n",
    "df_games = pd.read_csv(url_games, error_bad_lines=False)[['date', 'game', 'score', 'league', 'season', 'game_id', \n",
    "                'home', 'away', 'score_home', 'score_away', 'result']]\n",
    "\n",
    "print(df_games.shape[0], \" matches imported\")\n",
    "\n",
    "df_games.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18368  teamrows imported\n"
     ]
    },
    {
     "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>game_id</th>\n",
       "      <th>team</th>\n",
       "      <th>h_a</th>\n",
       "      <th>win_R5</th>\n",
       "      <th>draw_R5</th>\n",
       "      <th>lost_R5</th>\n",
       "      <th>H2H_W</th>\n",
       "      <th>H2H_D</th>\n",
       "      <th>H2H_L</th>\n",
       "      <th>scored_avg_R5</th>\n",
       "      <th>conceded_avg_R5</th>\n",
       "      <th>points_cum_prev_avg</th>\n",
       "      <th>points_cum_h_a_prev_avg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>422474</td>\n",
       "      <td>AIK</td>\n",
       "      <td>away</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>422481</td>\n",
       "      <td>AIK</td>\n",
       "      <td>home</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>422484</td>\n",
       "      <td>AIK</td>\n",
       "      <td>away</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.5</td>\n",
       "      <td>5.500000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>422494</td>\n",
       "      <td>AIK</td>\n",
       "      <td>home</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.333333</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>422499</td>\n",
       "      <td>AIK</td>\n",
       "      <td>away</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.5</td>\n",
       "      <td>3.750000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   game_id team   h_a    win_R5   draw_R5   lost_R5  H2H_W  H2H_D  H2H_L  \\\n",
       "0   422474  AIK  away  0.000000  0.000000  0.000000    0.0    0.0    0.0   \n",
       "1   422481  AIK  home  0.000000  0.000000  1.000000    0.0    0.0    0.0   \n",
       "2   422484  AIK  away  0.500000  0.000000  0.500000    0.0    0.0    0.0   \n",
       "3   422494  AIK  home  0.333333  0.333333  0.333333    0.0    0.0    0.0   \n",
       "4   422499  AIK  away  0.500000  0.250000  0.250000    0.0    0.0    0.0   \n",
       "\n",
       "   scored_avg_R5  conceded_avg_R5  points_cum_prev_avg  \\\n",
       "0            0.0         0.000000                  NaN   \n",
       "1            3.0         7.000000             0.000000   \n",
       "2            4.5         5.500000             1.500000   \n",
       "3            4.0         4.333333             1.666667   \n",
       "4            4.5         3.750000             2.000000   \n",
       "\n",
       "   points_cum_h_a_prev_avg  \n",
       "0                      NaN  \n",
       "1                      NaN  \n",
       "2                      0.0  \n",
       "3                      3.0  \n",
       "4                      1.0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Import team data\n",
    "url_teams = 'https://raw.githubusercontent.com/msjoelin/icehockey_analysis/master/data/df_teams.csv'\n",
    "\n",
    "df_teams = pd.read_csv(url_teams, error_bad_lines=False)[['game_id', 'team', 'h_a', \n",
    "                     'win_R5', 'draw_R5', 'lost_R5',  \n",
    "                     'H2H_W', 'H2H_D', 'H2H_L',\n",
    "                    'scored_avg_R5', 'conceded_avg_R5',\n",
    "                    'points_cum_prev_avg', 'points_cum_h_a_prev_avg']]\n",
    "\n",
    "print(df_teams.shape[0], \" teamrows imported\")\n",
    "\n",
    "df_teams.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Following columns were selected to be included in the analysis: \n",
    "\n",
    "* Win_R5 / Draw_R5 / Lost_R5   \n",
    "Ratio of won, draw or lost games the last 5 games\n",
    "* H2H_W/D/L  \n",
    "Head to head ratio\n",
    "* points_cum_prev_avg  \n",
    "Average number of points per game during the season so far\n",
    "* points_cum_h_a_prev_avg  \n",
    "Average number of points per game during the season so far when playing home or away\n",
    "\n",
    "Next step is to join data for home and away teams to the game dataset. These columns will be used for predictions later on. In order to do so, we are splitting up the teams data into home and away and adding suffix to column names. This data is then added to the games data into the new dataframe ```df_games_trend```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 9184 entries, 0 to 9183\n",
      "Data columns (total 28 columns):\n",
      " #   Column                        Non-Null Count  Dtype  \n",
      "---  ------                        --------------  -----  \n",
      " 0   date                          9184 non-null   object \n",
      " 1   game                          9184 non-null   object \n",
      " 2   score                         9184 non-null   object \n",
      " 3   league                        9184 non-null   object \n",
      " 4   season                        9184 non-null   object \n",
      " 5   game_id                       9184 non-null   int64  \n",
      " 6   home                          9184 non-null   object \n",
      " 7   away                          9184 non-null   object \n",
      " 8   score_home                    9184 non-null   int64  \n",
      " 9   score_away                    9184 non-null   int64  \n",
      " 10  result                        9184 non-null   object \n",
      " 11  win_R5_home                   9184 non-null   float64\n",
      " 12  draw_R5_home                  9184 non-null   float64\n",
      " 13  lost_R5_home                  9184 non-null   float64\n",
      " 14  H2H_W_home                    9184 non-null   float64\n",
      " 15  H2H_D_home                    9184 non-null   float64\n",
      " 16  H2H_L_home                    9184 non-null   float64\n",
      " 17  scored_avg_R5_home            9184 non-null   float64\n",
      " 18  conceded_avg_R5_home          9184 non-null   float64\n",
      " 19  points_cum_prev_avg_home      9009 non-null   float64\n",
      " 20  points_cum_h_a_prev_avg_home  8832 non-null   float64\n",
      " 21  win_R5_away                   9184 non-null   float64\n",
      " 22  draw_R5_away                  9184 non-null   float64\n",
      " 23  lost_R5_away                  9184 non-null   float64\n",
      " 24  scored_avg_R5_away            9184 non-null   float64\n",
      " 25  conceded_avg_R5_away          9184 non-null   float64\n",
      " 26  points_cum_prev_avg_away      9007 non-null   float64\n",
      " 27  points_cum_h_a_prev_avg_away  8832 non-null   float64\n",
      "dtypes: float64(17), int64(3), object(8)\n",
      "memory usage: 2.0+ MB\n"
     ]
    }
   ],
   "source": [
    "# Split into home and away data\n",
    "df_home = df_teams[df_teams['h_a']=='home'].copy().drop(['team', 'h_a'], axis=1).add_suffix('_home').rename(columns={\"game_id_home\": \"game_id\"})\n",
    "df_away = df_teams[df_teams['h_a']=='away'].copy().drop(['team', 'h_a'], axis=1).add_suffix('_away').rename(columns={\"game_id_away\": \"game_id\"})\n",
    "\n",
    "df_games_trend = pd.merge(df_games, df_home, \n",
    "                   on='game_id', how='left')\n",
    "\n",
    "df_games_trend = pd.merge(df_games_trend, df_away, \n",
    "                   on='game_id', how='left')\n",
    "\n",
    "# H2H values are now duplicates, so remove one of them and rename the other one\n",
    "df_games_trend=df_games_trend.drop(['H2H_W_away', 'H2H_D_away', 'H2H_L_away'], axis=1)\n",
    "\n",
    "df_games_trend.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That was it. We have now prepared the data that should be used for predictions. Next step is to explore the data. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2) Exploratory analysis\n",
    "\n",
    "As always the first step is to explore the data. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>game_id</th>\n",
       "      <th>score_home</th>\n",
       "      <th>score_away</th>\n",
       "      <th>win_R5_home</th>\n",
       "      <th>draw_R5_home</th>\n",
       "      <th>lost_R5_home</th>\n",
       "      <th>H2H_W_home</th>\n",
       "      <th>H2H_D_home</th>\n",
       "      <th>H2H_L_home</th>\n",
       "      <th>scored_avg_R5_home</th>\n",
       "      <th>conceded_avg_R5_home</th>\n",
       "      <th>points_cum_prev_avg_home</th>\n",
       "      <th>points_cum_h_a_prev_avg_home</th>\n",
       "      <th>win_R5_away</th>\n",
       "      <th>draw_R5_away</th>\n",
       "      <th>lost_R5_away</th>\n",
       "      <th>scored_avg_R5_away</th>\n",
       "      <th>conceded_avg_R5_away</th>\n",
       "      <th>points_cum_prev_avg_away</th>\n",
       "      <th>points_cum_h_a_prev_avg_away</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.00000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9009.000000</td>\n",
       "      <td>8832.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9184.000000</td>\n",
       "      <td>9007.000000</td>\n",
       "      <td>8832.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>256957.251089</td>\n",
       "      <td>2.935105</td>\n",
       "      <td>2.430205</td>\n",
       "      <td>0.375528</td>\n",
       "      <td>0.209776</td>\n",
       "      <td>0.395641</td>\n",
       "      <td>0.363284</td>\n",
       "      <td>0.202434</td>\n",
       "      <td>0.38169</td>\n",
       "      <td>2.603061</td>\n",
       "      <td>2.657571</td>\n",
       "      <td>1.467175</td>\n",
       "      <td>1.739727</td>\n",
       "      <td>0.393198</td>\n",
       "      <td>0.214315</td>\n",
       "      <td>0.373214</td>\n",
       "      <td>2.652134</td>\n",
       "      <td>2.596833</td>\n",
       "      <td>1.493979</td>\n",
       "      <td>1.218418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>141346.628142</td>\n",
       "      <td>1.717349</td>\n",
       "      <td>1.559005</td>\n",
       "      <td>0.242882</td>\n",
       "      <td>0.195283</td>\n",
       "      <td>0.245336</td>\n",
       "      <td>0.227754</td>\n",
       "      <td>0.174017</td>\n",
       "      <td>0.23556</td>\n",
       "      <td>0.899425</td>\n",
       "      <td>0.913987</td>\n",
       "      <td>0.473430</td>\n",
       "      <td>0.579692</td>\n",
       "      <td>0.246756</td>\n",
       "      <td>0.197301</td>\n",
       "      <td>0.241922</td>\n",
       "      <td>0.921188</td>\n",
       "      <td>0.894860</td>\n",
       "      <td>0.476068</td>\n",
       "      <td>0.566392</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>180.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>123022.750000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.230769</td>\n",
       "      <td>0.071429</td>\n",
       "      <td>0.25000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.181818</td>\n",
       "      <td>1.409091</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.200000</td>\n",
       "      <td>0.875000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>301008.500000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.37500</td>\n",
       "      <td>2.600000</td>\n",
       "      <td>2.600000</td>\n",
       "      <td>1.480769</td>\n",
       "      <td>1.750000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>2.600000</td>\n",
       "      <td>2.600000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.214286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>389831.250000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.50000</td>\n",
       "      <td>3.200000</td>\n",
       "      <td>3.200000</td>\n",
       "      <td>1.750000</td>\n",
       "      <td>2.090909</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.600000</td>\n",
       "      <td>3.200000</td>\n",
       "      <td>3.200000</td>\n",
       "      <td>1.764706</td>\n",
       "      <td>1.533333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>442238.000000</td>\n",
       "      <td>12.000000</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.00000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>7.500000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             game_id   score_home   score_away  win_R5_home  draw_R5_home  \\\n",
       "count    9184.000000  9184.000000  9184.000000  9184.000000   9184.000000   \n",
       "mean   256957.251089     2.935105     2.430205     0.375528      0.209776   \n",
       "std    141346.628142     1.717349     1.559005     0.242882      0.195283   \n",
       "min       180.000000     0.000000     0.000000     0.000000      0.000000   \n",
       "25%    123022.750000     2.000000     1.000000     0.200000      0.000000   \n",
       "50%    301008.500000     3.000000     2.000000     0.400000      0.200000   \n",
       "75%    389831.250000     4.000000     3.000000     0.600000      0.400000   \n",
       "max    442238.000000    12.000000    11.000000     1.000000      1.000000   \n",
       "\n",
       "       lost_R5_home   H2H_W_home   H2H_D_home  H2H_L_home  scored_avg_R5_home  \\\n",
       "count   9184.000000  9184.000000  9184.000000  9184.00000         9184.000000   \n",
       "mean       0.395641     0.363284     0.202434     0.38169            2.603061   \n",
       "std        0.245336     0.227754     0.174017     0.23556            0.899425   \n",
       "min        0.000000     0.000000     0.000000     0.00000            0.000000   \n",
       "25%        0.200000     0.230769     0.071429     0.25000            2.000000   \n",
       "50%        0.400000     0.363636     0.200000     0.37500            2.600000   \n",
       "75%        0.600000     0.500000     0.285714     0.50000            3.200000   \n",
       "max        1.000000     1.000000     1.000000     1.00000            8.000000   \n",
       "\n",
       "       conceded_avg_R5_home  points_cum_prev_avg_home  \\\n",
       "count           9184.000000               9009.000000   \n",
       "mean               2.657571                  1.467175   \n",
       "std                0.913987                  0.473430   \n",
       "min                0.000000                  0.000000   \n",
       "25%                2.000000                  1.181818   \n",
       "50%                2.600000                  1.480769   \n",
       "75%                3.200000                  1.750000   \n",
       "max               10.000000                  3.000000   \n",
       "\n",
       "       points_cum_h_a_prev_avg_home  win_R5_away  draw_R5_away  lost_R5_away  \\\n",
       "count                   8832.000000  9184.000000   9184.000000   9184.000000   \n",
       "mean                       1.739727     0.393198      0.214315      0.373214   \n",
       "std                        0.579692     0.246756      0.197301      0.241922   \n",
       "min                        0.000000     0.000000      0.000000      0.000000   \n",
       "25%                        1.409091     0.200000      0.000000      0.200000   \n",
       "50%                        1.750000     0.400000      0.200000      0.400000   \n",
       "75%                        2.090909     0.600000      0.400000      0.600000   \n",
       "max                        3.000000     1.000000      1.000000      1.000000   \n",
       "\n",
       "       scored_avg_R5_away  conceded_avg_R5_away  points_cum_prev_avg_away  \\\n",
       "count         9184.000000           9184.000000               9007.000000   \n",
       "mean             2.652134              2.596833                  1.493979   \n",
       "std              0.921188              0.894860                  0.476068   \n",
       "min              0.000000              0.000000                  0.000000   \n",
       "25%              2.000000              2.000000                  1.200000   \n",
       "50%              2.600000              2.600000                  1.500000   \n",
       "75%              3.200000              3.200000                  1.764706   \n",
       "max             10.000000              7.500000                  3.000000   \n",
       "\n",
       "       points_cum_h_a_prev_avg_away  \n",
       "count                   8832.000000  \n",
       "mean                       1.218418  \n",
       "std                        0.566392  \n",
       "min                        0.000000  \n",
       "25%                        0.875000  \n",
       "50%                        1.214286  \n",
       "75%                        1.533333  \n",
       "max                        3.000000  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_games_trend.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset contains from 2007 for SHL and Allsvenskan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2015/16    728\n",
      "2019/20    728\n",
      "2017/18    728\n",
      "2018/19    728\n",
      "2016/17    728\n",
      "2010/11    694\n",
      "2014/15    694\n",
      "2013/14    694\n",
      "2011/12    694\n",
      "2009/10    694\n",
      "2012/13    694\n",
      "2007/08    690\n",
      "2008/09    690\n",
      "Name: season, dtype: int64\n",
      "allsvenskan    4724\n",
      "shl            4460\n",
      "Name: league, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(df_games_trend['season'].value_counts())\n",
    "print(df_games_trend['league'].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 20 entries, game_id to points_cum_h_a_prev_avg_away\n",
      "Data columns (total 20 columns):\n",
      " #   Column                        Non-Null Count  Dtype  \n",
      "---  ------                        --------------  -----  \n",
      " 0   game_id                       20 non-null     float64\n",
      " 1   score_home                    20 non-null     float64\n",
      " 2   score_away                    20 non-null     float64\n",
      " 3   win_R5_home                   20 non-null     float64\n",
      " 4   draw_R5_home                  20 non-null     float64\n",
      " 5   lost_R5_home                  20 non-null     float64\n",
      " 6   H2H_W_home                    20 non-null     float64\n",
      " 7   H2H_D_home                    20 non-null     float64\n",
      " 8   H2H_L_home                    20 non-null     float64\n",
      " 9   scored_avg_R5_home            20 non-null     float64\n",
      " 10  conceded_avg_R5_home          20 non-null     float64\n",
      " 11  points_cum_prev_avg_home      20 non-null     float64\n",
      " 12  points_cum_h_a_prev_avg_home  20 non-null     float64\n",
      " 13  win_R5_away                   20 non-null     float64\n",
      " 14  draw_R5_away                  20 non-null     float64\n",
      " 15  lost_R5_away                  20 non-null     float64\n",
      " 16  scored_avg_R5_away            20 non-null     float64\n",
      " 17  conceded_avg_R5_away          20 non-null     float64\n",
      " 18  points_cum_prev_avg_away      20 non-null     float64\n",
      " 19  points_cum_h_a_prev_avg_away  20 non-null     float64\n",
      "dtypes: float64(20)\n",
      "memory usage: 3.3+ KB\n"
     ]
    }
   ],
   "source": [
    "df_games_trend.corr().info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next step is to add some variables in order to analyze correlation of outcome vs attributes. \n",
    "There will be three new columns, each having 1 or 0 if the outcome was home, draw or away. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Introduce binary outcome variables\n",
    "df_games_trend['home_result'] = 0\n",
    "df_games_trend.loc[(df_games_trend['result'] == 'home'), 'home_result'] = 1\n",
    "\n",
    "df_games_trend['draw_result'] = 0\n",
    "df_games_trend.loc[(df_games_trend['result'] == 'draw'), 'draw_result'] = 1\n",
    "\n",
    "df_games_trend['away_result'] = 0\n",
    "df_games_trend.loc[(df_games_trend['result'] == 'away'), 'away_result'] = 1\n",
    "\n",
    "# Heatmap for correlations\n",
    "plt.figure(figsize = (12, 9))\n",
    "s = sns.heatmap(df_games_trend.corr()[['home_result', 'draw_result', 'away_result']][3:20],\n",
    "               annot = True, \n",
    "               cmap = 'RdBu',\n",
    "               vmin = -0.1, \n",
    "               vmax = 0.1)\n",
    "s.set_yticklabels(s.get_yticklabels(), rotation = 0, fontsize = 12)\n",
    "s.set_xticklabels(s.get_xticklabels(), rotation = 90, fontsize = 12)\n",
    "plt.title('Correlation Heatmap')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The heatmap shows correlation between result and different attributes. \n",
    "\n",
    "* Highest correlation for home and away win is the total cumulative points in the season. Not very surprising. \n",
    "* Importance of form for the home team has higher correlation than for away team. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Prepare data for the machine learning model\n",
    "\n",
    "After the quick cleaning and exploration of the data, it's time to set up the machine learning model.  \n",
    "Out dataframe contains some null values, coming from games in the beginning of the season where no trends can be calculated. These are dropped, together with the columns that was created for the correlation chart. \n",
    "\n",
    "The data for current season is saved to another dataframe, as this will be used to evaluate the final performance of the model and thus should not be used when training the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data current season: 697\n",
      "Data for modeling: 8107\n"
     ]
    }
   ],
   "source": [
    "df_model = df_games_trend[~df_games_trend.isnull().any(axis=1)].reset_index(drop=True)\n",
    "df_model  = df_model.drop(['home_result','draw_result', 'away_result'], axis=1)\n",
    "\n",
    "df_2019 = df_model[df_model['season']=='2019/20']\n",
    "df_model = df_model[df_model['season']!='2019/20'].reset_index(drop=True)\n",
    "\n",
    "print(\"Data current season: \" + str(df_2019.shape[0]))\n",
    "print(\"Data for modeling: \" + str(df_model.shape[0]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check for imbalance\n",
    "A balanced dataset would give us about 1/3 per each home, away and draw outcome.  \n",
    "\n",
    "Now we have a skewed distribution, and thus a baseline, primitive model only predicting \"home\" all the time would be right 50% of the times. This is important to keep in mind when evaluating the final outcome. The trained model should at least be better than this. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "home    0.472555\n",
       "away    0.308869\n",
       "draw    0.218577\n",
       "Name: result, dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_model['result'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Prepare data for the model\n",
    "\n",
    "This is a classification problem. Random forest is in general a popular algorithm to use. Random forest is kind of an advanced decision tree. The classifier from sklearn will be used. For this to work, the classes must be turned into factors. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "#Creating the dependent variable class\n",
    "mapping = {'home': 0, 'draw': 1, 'away': 2}\n",
    "df_model = df_model.replace({'result': mapping})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Split into dependent and independent variables\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 8107 entries, 0 to 8106\n",
      "Data columns (total 28 columns):\n",
      " #   Column                        Non-Null Count  Dtype  \n",
      "---  ------                        --------------  -----  \n",
      " 0   date                          8107 non-null   object \n",
      " 1   game                          8107 non-null   object \n",
      " 2   score                         8107 non-null   object \n",
      " 3   league                        8107 non-null   object \n",
      " 4   season                        8107 non-null   object \n",
      " 5   game_id                       8107 non-null   int64  \n",
      " 6   home                          8107 non-null   object \n",
      " 7   away                          8107 non-null   object \n",
      " 8   score_home                    8107 non-null   int64  \n",
      " 9   score_away                    8107 non-null   int64  \n",
      " 10  result                        8107 non-null   int64  \n",
      " 11  win_R5_home                   8107 non-null   float64\n",
      " 12  draw_R5_home                  8107 non-null   float64\n",
      " 13  lost_R5_home                  8107 non-null   float64\n",
      " 14  H2H_W_home                    8107 non-null   float64\n",
      " 15  H2H_D_home                    8107 non-null   float64\n",
      " 16  H2H_L_home                    8107 non-null   float64\n",
      " 17  scored_avg_R5_home            8107 non-null   float64\n",
      " 18  conceded_avg_R5_home          8107 non-null   float64\n",
      " 19  points_cum_prev_avg_home      8107 non-null   float64\n",
      " 20  points_cum_h_a_prev_avg_home  8107 non-null   float64\n",
      " 21  win_R5_away                   8107 non-null   float64\n",
      " 22  draw_R5_away                  8107 non-null   float64\n",
      " 23  lost_R5_away                  8107 non-null   float64\n",
      " 24  scored_avg_R5_away            8107 non-null   float64\n",
      " 25  conceded_avg_R5_away          8107 non-null   float64\n",
      " 26  points_cum_prev_avg_away      8107 non-null   float64\n",
      " 27  points_cum_h_a_prev_avg_away  8107 non-null   float64\n",
      "dtypes: float64(17), int64(4), object(7)\n",
      "memory usage: 1.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df_model.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dependent variable (what should be predicted) is in column 10, and the independent variables are in columns 11 to 27. \n",
    "A third variable with attributes are stored in the variable ```info```."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Splitting the data into independent and dependent variables\n",
    "X = df_model.iloc[:,11:].values\n",
    "y = df_model.iloc[:,10].values\n",
    "info = df_model.iloc[:,0:9].values\n",
    "\n",
    "X_2019 = df_2019.iloc[:,11:].values\n",
    "y_2019 = df_2019.iloc[:,10].values\n",
    "info_2019 = df_2019.iloc[:,0:9].values\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data is split into train and test data, splitted 80/20.  \n",
    "Considering our imbalanced data set, the input ``stratify =y`` is used in the function.  \n",
    "\n",
    "Finally the independent variables has to be scaled as they are of different scales. This is done with the standard scaler in sklearn library. The mean value for each variable is subtracted and then divided by its unit variance . "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Random state = 21 (random number)\n",
    "X_train, X_test, y_train, y_test, info_train, info_test  = train_test_split(X, y, info, \n",
    "                                                                            test_size = 0.2, stratify = y, random_state = 21)\n",
    "\n",
    "# Feature Scaling\n",
    "scaler = StandardScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)\n",
    "X_2019 = scaler.transform(X_2019)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Train model\n",
    "\n",
    "Data is ready to be used to train the model. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'RandomForestClassifier' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-1-a6e13b26891c>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;31m# Fitting Random Forest Classification to the Training set\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mrf\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mRandomForestClassifier\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mn_estimators\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m100\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mrandom_state\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;36m42\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m \u001b[0mrf\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0my_train\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m \u001b[1;31m# save the model to disk\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'RandomForestClassifier' is not defined"
     ]
    }
   ],
   "source": [
    "# Fitting Random Forest Classification to the Training set\n",
    "rf = RandomForestClassifier(n_estimators = 100, criterion = 'entropy', random_state = 42)\n",
    "rf.fit(X_train, y_train)\n",
    "\n",
    "# save the model to disk\n",
    "#filename = \"C:/Users/marcu/Documents/projects/icehockey/ml_models/rf_predict_game.sav\"\n",
    "#pickle.dump(rf, open(filename, 'wb'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Calculate feature importances\n",
    "importances = rf.feature_importances_\n",
    "# Sort feature importances in descending order\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "names = ['win_R5_home', 'draw_R5_home', 'lost_R5_home', \n",
    "         'H2H_W_home', 'H2H_D_home', 'H2H_L_home', \n",
    "         'scored_avg_R5_home', 'conceded_avg_R5_home',\n",
    "         'points_cum_prev_avg_home', 'points_cum_h_a_prev_avg_home',\n",
    "         'win_R5_away', 'draw_R5_away', 'lost_R5_away',\n",
    "         'scored_avg_R5_away', 'conceded_avg_R5_away',\n",
    "         'points_cum_prev_avg_away', 'points_cum_h_a_prev_avg_away']\n",
    "\n",
    "\n",
    "# Barplot: Add bars\n",
    "plt.bar(range(X.shape[1]), importances[indices])\n",
    "# Add feature names as x-axis labels\n",
    "plt.xticks(range(X.shape[1]), names, rotation=30, fontsize = 8)\n",
    "# Create plot title\n",
    "plt.title(\"Feature Importance\")\n",
    "# Show plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.DataFrame({'val': importances}, index=names)\n",
    "ax = df.plot.barh(y='val')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5. Evaluate performance\n",
    "\n",
    "Time to evaluate the performance of the model. This is done with precision and recall:  \n",
    "\n",
    "* Precision: Proportion of positive predictions that was correct\n",
    "* Recall: Proportion of actual positives that was identified"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.81      0.62       766\n",
      "           1       0.27      0.03      0.06       355\n",
      "           2       0.42      0.28      0.33       501\n",
      "\n",
      "    accuracy                           0.48      1622\n",
      "   macro avg       0.40      0.37      0.34      1622\n",
      "weighted avg       0.42      0.48      0.41      1622\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred_rf = rf.predict(X_test)\n",
    "y_pred_rf_prob = rf.predict_proba(X_test)\n",
    "\n",
    "print(classification_report(y_test, y_pred_rf))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* If the model predicts 0 (home win), this is correct 1/2 of the times. The model identifies 81% of all home wins.  \n",
    "* If the model predicts 1 (draw), this is only correct in 27% of the cases. The model has a hard time to identify these (only 3 % are discovered) \n",
    "* If the model predicts 2 (away), this is correct in 42 % of the cases. 28 % of the away wins are identified.\n",
    "\n",
    "The reason for low numbers for draw is that this sign is the least frequent. Thus, the model tends to choose either home or away wins to a higher extent. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Would using the model current season bring any money?\n",
    "\n",
    "Now to the interesting part: Would the model bring any money?\n",
    "We are testing the model on the data for 2019. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred2019_rf = rf.predict(X_2019)\n",
    "y_pred2019_rf_prob = rf.predict_proba(X_2019)\n",
    "\n",
    "df_2019 = pd.DataFrame(info_2019)\n",
    "\n",
    "df_2019.columns = ['date', 'game', 'score', 'league', 'season', 'game_id', 'home', 'away','score_home']\n",
    "\n",
    "# Set the outcome variables and probabilities\n",
    "df_2019['actual'] = y_2019\n",
    "df_2019['predicted'] = y_pred2019_rf\n",
    "df_2019['prob_home'] = y_pred2019_rf_prob[:,0]\n",
    "df_2019['prob_draw'] = y_pred2019_rf_prob[:,1]\n",
    "df_2019['prob_away'] = y_pred2019_rf_prob[:,2]\n",
    "\n",
    "# Map back sign outcome\n",
    "mapping = {0:'home', 1:'draw', 2:'away'}\n",
    "df_2019 = df_2019.replace({'actual': mapping, 'predicted': mapping})\n",
    "\n",
    "df_2019['correct']=0\n",
    "df_2019.loc[df_2019['actual']==df_2019['predicted'],'correct']=1\n",
    "\n",
    "# Check mean values by sign\n",
    "print(\"Correct ratio by prediction\")\n",
    "df_2019.groupby(['predicted'])['correct'].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The accuracy looks pretty good. \n",
    "For a proper evaluation, we need the odds per outcome. This data has been collected with the notebook \n",
    "https://github.com/msjoelin/icehockey_analysis/blob/master/get_odds_liverattning.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add odds data to games\n",
    "url_odds = 'https://raw.githubusercontent.com/msjoelin/icehockey_analysis/master/data/sweden_odds.csv'\n",
    "\n",
    "df_odds = pd.read_csv(url_odds, error_bad_lines=False, sep=',', decimal=',')\n",
    "\n",
    "df_2019_odds = pd.merge(df_2019, df_odds[['date', 'home', 'away', 'odds_1', 'odds_X', 'odds_2', 'odds_outcome']],\n",
    "                   on=['date', 'home', 'away'], how='left')\n",
    "\n",
    "# Add columns for return\n",
    "df_2019_odds['return'] = df_2019_odds['odds_outcome']*df_2019_odds['correct']\n",
    "df_2019_odds.loc[df_2019_odds['correct']==0,'return'] = -1\n",
    "\n",
    "df_2019_odds.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.barplot(x='predicted', y='return', \n",
    "            estimator = sum, data=df_2019_odds, ci=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_2019_odds.to_csv(\"C:/Users/marcu/Documents/projects/icehockey/data/df_2019_odds.csv\", index=False)"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}