Pitch_Predictor_1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normalized confusion matrix\n",
      "[[0.56260494 0.2314761  0.03782806 0.         0.14153027 0.02656063\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.42639449 0.36964562 0.04594676 0.         0.1398229  0.01819022\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.46641193 0.19023151 0.19360159 0.         0.12622373 0.02353123\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.4617037  0.22864218 0.0823391  0.         0.19828152 0.02903349\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.26711028 0.09122732 0.01015471 0.         0.60984429 0.02166339\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.37802597 0.24242112 0.05884674 0.         0.2518915  0.06881466\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.34933051 0.14132308 0.04650669 0.         0.40906801 0.05377171\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.36353699 0.19282157 0.05064982 0.         0.31926394 0.07372768\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.31645455 0.24767578 0.03935373 0.         0.32177994 0.07473599\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.16154118 0.01343231 0.00813008 0.         0.64581124 0.17108519\n",
      "  0.         0.         0.         0.         0.        ]\n",
      " [0.19444444 0.00609756 0.00745257 0.         0.79200542 0.\n",
      "  0.         0.         0.         0.         0.        ]]\n",
      "program took 1594.6933572292328 seconds to complete!\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       0.44      0.56      0.49    253119\n",
      "          6       0.25      0.37      0.30    112478\n",
      "          7       0.37      0.19      0.25     84271\n",
      "          5       0.00      0.00      0.00     73088\n",
      "          4       0.24      0.61      0.34     60563\n",
      "          3       0.18      0.07      0.10     58287\n",
      "         11       0.00      0.00      0.00     37715\n",
      "          2       0.00      0.00      0.00     16466\n",
      "          8       0.00      0.00      0.00     11079\n",
      "          9       0.00      0.00      0.00      2829\n",
      "         10       0.00      0.00      0.00      1476\n",
      "\n",
      "avg / total       0.27      0.34      0.29    711371\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1135: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          1       0.44      0.56      0.49    760092\n",
      "          6       0.25      0.37      0.30    337174\n",
      "          7       0.37      0.19      0.26    253090\n",
      "          5       0.00      0.00      0.00    219347\n",
      "          4       0.24      0.61      0.34    181469\n",
      "          3       0.18      0.07      0.10    175692\n",
      "         11       0.00      0.00      0.00    111749\n",
      "          2       0.00      0.00      0.00     49896\n",
      "          8       0.00      0.00      0.00     32567\n",
      "          9       0.00      0.00      0.00      8415\n",
      "         10       0.00      0.00      0.00      4621\n",
      "\n",
      "avg / total       0.27      0.34      0.29   2134112\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "np.random.seed(0)\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import BaggingClassifier, RandomForestClassifier\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import roc_auc_score, roc_curve, auc\n",
    "from sklearn.metrics import classification_report\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_style('darkgrid')\n",
    "%matplotlib inline\n",
    "from imblearn.over_sampling import SMOTE, ADASYN\n",
    "from plot_confusion_matrix import *\n",
    "import time\n",
    "\n",
    "# step 1. imported all data\n",
    "start_time = time.time()\n",
    "\n",
    "df_batter = pd.read_csv('atbats.csv')\n",
    "df_names = pd.read_csv('player_names.csv')\n",
    "df_names.columns = [\"pitcher_id\", \"first_name\", \"last_name\"]\n",
    "df_batter_pitcher= df_batter.merge(df_names, on = 'pitcher_id', how = 'left')\n",
    "df_pitches = pd.read_csv('pitches.csv')\n",
    "df_full= df_batter_pitcher.merge(df_pitches, on = 'ab_id', how = 'left')\n",
    "\n",
    "# step 2. created a dummy variable\n",
    "dummy = pd.get_dummies(df_full['type'])\n",
    "df_full = df_full.merge(dummy, left_index = True, right_index = True)\n",
    "\n",
    "# step 3. rename columns\n",
    "df_full.rename(columns =({'p_score': 'Pitcher_Score'}),inplace = True)\n",
    "df_full.rename(columns =({'b_score': 'Batter_Score'}),inplace = True)\n",
    "df_full.rename(columns =({'b_count': 'Ball_Count'}),inplace = True)\n",
    "df_full.rename(columns =({'s_count': 'Strike_Count'}),inplace = True)\n",
    "\n",
    "# step 4. drop columns\n",
    "column_names = ['ab_id','p_throws', 'stand', 'type', 'event', 'g_id', 'o', 'top', 'first_name', 'last_name',\n",
    "       'px', 'pz', 'start_speed', 'end_speed', 'spin_rate', 'spin_dir',\n",
    "       'break_angle', 'break_length', 'break_y', 'ax', 'ay', 'az', 'sz_bot',\n",
    "       'sz_top', 'type_confidence', 'vx0', 'vy0', 'vz0', 'x', 'x0', 'y', 'y0',\n",
    "       'z0', 'pfx_x', 'pfx_z', 'nasty', 'zone', 'code',\n",
    "       'event_num', 'pitch_num']\n",
    "df_full = df_full.drop(columns=column_names, axis=1)\n",
    "\n",
    "# step 5. drop rows\n",
    "df_full = df_full[(df_full['pitch_type'] != 'FA') & \n",
    "                  (df_full['pitch_type'] != 'AB') &\n",
    "                  (df_full['pitch_type'] != 'UN') &\n",
    "                  (df_full['pitch_type'] != 'SC') &\n",
    "                  (df_full['pitch_type'] != 'FO') &\n",
    "                  (df_full['pitch_type'] != 'PO') &\n",
    "                  (df_full['pitch_type'] != 'EP')]\n",
    "\n",
    "# step 6. add prev_pitch\n",
    "df_full['prev_pitch'] = df_full['pitch_type'].shift(1)\n",
    "df_full.dropna(inplace=True)\n",
    "\n",
    "# step 7. convert categorical variable into numeric variable\n",
    "score_to_numeric = {\n",
    "    \"FF\" : 1,\n",
    "    \"SL\" : 2,\n",
    "    \"FT\" : 3,\n",
    "    \"CH\" : 4,\n",
    "    \"SI\" : 5,\n",
    "    \"CU\" : 6,\n",
    "    \"FC\" : 7,\n",
    "    \"KC\" : 8,\n",
    "    \"FS\" : 9,\n",
    "    \"KN\" : 10,\n",
    "    \"IN\" : 11\n",
    "}\n",
    "\n",
    "df_full['prev_pitch'] = df_full['prev_pitch'].apply(lambda x : score_to_numeric[x])\n",
    "df_full['pitch_type'] = df_full['pitch_type'].apply(lambda x : score_to_numeric[x])\n",
    "\n",
    "# step 9. prepare for training\n",
    "\n",
    "target = df_full['pitch_type']\n",
    "df_full.drop('pitch_type', axis = 1, inplace = True)\n",
    "df_full_copy = df_full\n",
    "df_full_copy = df_full_copy.drop(columns=[\"pitcher_id\", \"batter_id\"])\n",
    "\n",
    "data_train, data_test, target_train, target_test = train_test_split(df_full_copy,\n",
    "                                                                    target,\n",
    "                                                                    test_size = 0.25, \n",
    "                                                                    random_state=123)\n",
    "\n",
    "\n",
    "# step 10. training\n",
    "sampling_matrix = {\n",
    "    1: 0.05,\n",
    "    2: 0.1,\n",
    "    3: 0.1,\n",
    "    4: 0.1,\n",
    "    5: 0.125,\n",
    "    6: 0.125,\n",
    "    7: 0.1,\n",
    "    8: 0.075,\n",
    "    9: 0.075,\n",
    "    10: 0.075,\n",
    "    11: 0.075\n",
    "}\n",
    "\n",
    "forest = RandomForestClassifier(class_weight=sampling_matrix,\n",
    "                                n_estimators=300, \n",
    "                                max_depth=4)\n",
    "\n",
    "forest.fit(data_train, target_train)\n",
    "\n",
    "# step 11. predictions\n",
    "y_test = forest.predict(data_test)\n",
    "y_train = forest.predict(data_train)\n",
    "target_list = list(target.unique())\n",
    "\n",
    "# step 12. visualizations\n",
    "\n",
    "def plot_confusion_matrix(y_true, y_pred, classes,\n",
    "                          normalize=False,\n",
    "                          title=None,\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if not title:\n",
    "        if normalize:\n",
    "            title = 'Normalized confusion matrix'\n",
    "        else:\n",
    "            title = 'Confusion matrix, without normalization'\n",
    "\n",
    "    # Compute confusion matrix\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    # Only use the labels that appear in the data\n",
    "    # classes = np.array(classes)[unique_labels(np.array(y_true), y_pred)]\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    # We want to show all ticks...\n",
    "    ax.set(xticks=np.arange(cm.shape[1]),\n",
    "           yticks=np.arange(cm.shape[0]),\n",
    "           # ... and label them with the respective list entries\n",
    "           xticklabels=classes, yticklabels=classes,\n",
    "           title=title,\n",
    "           ylabel='True label',\n",
    "           xlabel='Predicted label')\n",
    "\n",
    "    # Rotate the tick labels and set their alignment.\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",\n",
    "             rotation_mode=\"anchor\")\n",
    "\n",
    "    # Loop over data dimensions and create text annotations.\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(j, i, format(cm[i, j], fmt),\n",
    "                    ha=\"center\", va=\"center\",\n",
    "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    fig.tight_layout()\n",
    "    return ax\n",
    "\n",
    "plot_confusion_matrix(np.array(target_test), \n",
    "                      y_test,\n",
    "                      target_list,\n",
    "                      normalize=True,\n",
    "                      title=None,\n",
    "                      cmap=plt.cm.Blues)\n",
    "\n",
    "\n",
    "print(f\"program took {time.time() - start_time} seconds to complete!\")\n",
    "\n",
    "#step 13. classification reports\n",
    "print(classification_report(list(target_test), list(y_test), target_names=[str(i) for i in target_list]))\n",
    "print(classification_report(list(target_train), list(y_train), target_names=[str(i) for i in target_list]))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "learn-env",
   "language": "python",
   "name": "learn-env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}