GB Modeling Notebook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ed224a17",
   "metadata": {},
   "outputs": [],
   "source": [
    "from src import Modeler, Processor, Reader\n",
    "import pandas as pd\n",
    "from datetime import datetime\n",
    "from sklearn import metrics, model_selection, preprocessing, inspection\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "import numpy as np\n",
    "import time\n",
    "\n",
    "# Plotting\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns \n",
    "sns.set()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7b5c5254",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Using the US Census regions from here: https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf\n",
    "south_atlantic_states = [\n",
    "    10   # Delaware\n",
    "    , 11 # DC\n",
    "    , 12 # Florida\n",
    "    , 13 # Georgia\n",
    "    , 24 # Maryland\n",
    "    , 37 # North Carolina\n",
    "    , 45 # South Carolina\n",
    "    , 51 # Virginia\n",
    "    , 54 # West Virginia\n",
    "]\n",
    "state_filter = south_atlantic_states"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7eb214ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "state_dict = {\n",
    "    'code' : ['10', '11', '12', '13', '24', '37', '45', '51', '54']\n",
    "    , 'state_name' : ['Delware', 'DC', 'Florida', 'Georgia', 'Maryland', 'North Carolina', 'South Carolina', 'Virginia', 'West Virginia']\n",
    "}\n",
    "\n",
    "state_code_map = pd.DataFrame(data=state_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23996f4c",
   "metadata": {},
   "source": [
    "## Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b6d8b615",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Instantiate preprocessor\n",
    "preprocessor = Processor.PreProcessor(state_filter=state_filter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "240f8773",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/elizabethmartens/cs5644/project/projectGitRepo/src/Processor.py:37: DtypeWarning: Columns (15) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  merged_df = self.init_cdc_data().merge(self.init_vaccinations_data(), how='left', on=['date', 'FIPS'])\n",
      "/Users/elizabethmartens/cs5644/project/projectGitRepo/src/Processor.py:37: DtypeWarning: Columns (15) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  merged_df = self.init_cdc_data().merge(self.init_vaccinations_data(), how='left', on=['date', 'FIPS'])\n"
     ]
    }
   ],
   "source": [
    "# Put data into a df to work with\n",
    "df = preprocessor.get_processed_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9976e6a4",
   "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>FIPS_10001</th>\n",
       "      <th>FIPS_10003</th>\n",
       "      <th>FIPS_10005</th>\n",
       "      <th>FIPS_11001</th>\n",
       "      <th>FIPS_12001</th>\n",
       "      <th>FIPS_12003</th>\n",
       "      <th>FIPS_12005</th>\n",
       "      <th>FIPS_12007</th>\n",
       "      <th>FIPS_12009</th>\n",
       "      <th>FIPS_12011</th>\n",
       "      <th>...</th>\n",
       "      <th>is_metro</th>\n",
       "      <th>Series_Complete_Pop_Pct_UR_Equity</th>\n",
       "      <th>cases</th>\n",
       "      <th>deaths</th>\n",
       "      <th>8_days_prior_cases</th>\n",
       "      <th>9_days_prior_cases</th>\n",
       "      <th>10_days_prior_cases</th>\n",
       "      <th>8_days_prior_deaths</th>\n",
       "      <th>9_days_prior_deaths</th>\n",
       "      <th>10_days_prior_deaths</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>25.142857</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>3.428571</td>\n",
       "      <td>3.857143</td>\n",
       "      <td>3.428571</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.714286</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>4.714286</td>\n",
       "      <td>3.428571</td>\n",
       "      <td>3.857143</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>27.857143</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>6.142857</td>\n",
       "      <td>4.714286</td>\n",
       "      <td>3.428571</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>22.857143</td>\n",
       "      <td>0.428571</td>\n",
       "      <td>8.714286</td>\n",
       "      <td>6.142857</td>\n",
       "      <td>4.714286</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.285714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>14.428571</td>\n",
       "      <td>8.714286</td>\n",
       "      <td>6.142857</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 704 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  FIPS_10001 FIPS_10003 FIPS_10005 FIPS_11001 FIPS_12001 FIPS_12003  \\\n",
       "0        1.0        0.0        0.0        0.0        0.0        0.0   \n",
       "1        1.0        0.0        0.0        0.0        0.0        0.0   \n",
       "2        1.0        0.0        0.0        0.0        0.0        0.0   \n",
       "3        1.0        0.0        0.0        0.0        0.0        0.0   \n",
       "4        1.0        0.0        0.0        0.0        0.0        0.0   \n",
       "\n",
       "  FIPS_12005 FIPS_12007 FIPS_12009 FIPS_12011  ... is_metro  \\\n",
       "0        0.0        0.0        0.0        0.0  ...      1.0   \n",
       "1        0.0        0.0        0.0        0.0  ...      1.0   \n",
       "2        0.0        0.0        0.0        0.0  ...      1.0   \n",
       "3        0.0        0.0        0.0        0.0  ...      1.0   \n",
       "4        0.0        0.0        0.0        0.0  ...      1.0   \n",
       "\n",
       "  Series_Complete_Pop_Pct_UR_Equity      cases    deaths 8_days_prior_cases  \\\n",
       "0                               0.0  25.142857  0.142857           3.428571   \n",
       "1                               0.0  26.714286  0.285714           4.714286   \n",
       "2                               0.0  27.857143  0.285714           6.142857   \n",
       "3                               0.0  22.857143  0.428571           8.714286   \n",
       "4                               0.0       26.0  0.571429          14.428571   \n",
       "\n",
       "  9_days_prior_cases 10_days_prior_cases 8_days_prior_deaths  \\\n",
       "0           3.857143            3.428571            0.285714   \n",
       "1           3.428571            3.857143            0.285714   \n",
       "2           4.714286            3.428571                 0.0   \n",
       "3           6.142857            4.714286                 0.0   \n",
       "4           8.714286            6.142857            0.142857   \n",
       "\n",
       "  9_days_prior_deaths 10_days_prior_deaths  \n",
       "0            0.285714             0.285714  \n",
       "1            0.285714             0.285714  \n",
       "2            0.285714             0.285714  \n",
       "3                 0.0             0.285714  \n",
       "4                 0.0                  0.0  \n",
       "\n",
       "[5 rows x 704 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Export data to csv so that the data does not have to be recreated each run\n",
    "df.to_csv('COVID_Analysis_Data.csv',index=False)\n",
    "# Show data\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cb9b616c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "#df = pd.read_csv(\"COVID_Analysis_Data.csv\", dtypes={'date':''})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "73a30dd4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(289884, 704)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "118b1a8c",
   "metadata": {},
   "source": [
    "## Split into train and test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "44debb1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "modeler = Modeler.Modeler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ff1bf4b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split data into training and test dataset:\n",
    "train_end_date = '2021-06-14'\n",
    "X_train, y_train_cases, y_train_deaths, X_test, y_test_cases, y_test_deaths = modeler.split_train_test(df, train_end_date)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52ab2d10",
   "metadata": {},
   "source": [
    "### Unpivoted data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7e16c7bc",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/elizabethmartens/cs5644/project/projectGitRepo/src/Processor.py:37: DtypeWarning: Columns (15) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  merged_df = self.init_cdc_data().merge(self.init_vaccinations_data(), how='left', on=['date', 'FIPS'])\n",
      "/Users/elizabethmartens/cs5644/project/projectGitRepo/src/Processor.py:37: DtypeWarning: Columns (15) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  merged_df = self.init_cdc_data().merge(self.init_vaccinations_data(), how='left', on=['date', 'FIPS'])\n"
     ]
    }
   ],
   "source": [
    "unpivoted_df = preprocessor.get_processed_data_without_fips_as_columns(set_current_data=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cf6707be",
   "metadata": {},
   "outputs": [],
   "source": [
    "unpivoted_cases = unpivoted_df[['date', 'FIPS', 'cases']].copy()\n",
    "unpivoted_cases['STATE'] = unpivoted_cases['FIPS'].apply(lambda x: x[:2])\n",
    "unpivoted_cases = unpivoted_cases.merge(state_code_map, left_on='STATE', right_on='code', how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "102b9a7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "unpivoted_cases_test = unpivoted_cases[unpivoted_cases['date'] > train_end_date]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "bf2e0d08",
   "metadata": {},
   "outputs": [],
   "source": [
    "unpivoted_deaths = unpivoted_df[['date', 'FIPS', 'deaths']].copy()\n",
    "unpivoted_deaths['STATE'] = unpivoted_deaths['FIPS'].apply(lambda x: x[:2])\n",
    "unpivoted_deaths = unpivoted_deaths.merge(state_code_map, left_on='STATE', right_on='code', how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "a4c535d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "unpivoted_deaths_test = unpivoted_deaths[unpivoted_deaths['date'] > train_end_date]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a9968ceb",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "13a39dd3",
   "metadata": {},
   "source": [
    "## Plotting functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5f4ff563",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_predictions(y_test, y_pred, model_name=None):\n",
    "    plt.figure(figsize=(9,6))\n",
    "    plt.scatter(y_test, y_pred, color='none', edgecolor='b')\n",
    "    xymin = min(np.min(y_test), np.min(y_pred))\n",
    "    xymax = max(np.max(y_test), np.max(y_pred))\n",
    "    plt.plot([xymin, xymax],[xymin, xymax], color=\"r\", linestyle=\"--\")\n",
    "    plt.xlabel('Actual')\n",
    "    plt.ylabel('Predicted')\n",
    "    if model_name:\n",
    "        plt.title(f'{model_name}: Actuals vs Predicted', fontsize=16)\n",
    "    else:\n",
    "        plt.title('Actuals vs Predicted', fontsize=16)\n",
    "    plt.show()\n",
    "    \n",
    "def plot_predictions_w_state(y_test, y_pred, unpivoted_test_df, model_name=None):\n",
    "    plt.figure(figsize=(9,6))\n",
    "    sns.scatterplot(x=y_test, y=y_pred, hue=unpivoted_test_df['state_name'])\n",
    "    xymin = min(np.min(y_test), np.min(y_pred))\n",
    "    xymax = max(np.max(y_test), np.max(y_pred))\n",
    "    plt.plot([xymin, xymax],[xymin, xymax], color=\"r\", linestyle=\"--\")\n",
    "    plt.xlabel('Actual')\n",
    "    plt.ylabel('Predicted')\n",
    "    if model_name:\n",
    "        plt.title(f'{model_name}: Actuals vs Predicted by state', fontsize=16)\n",
    "    else:\n",
    "        plt.title('Actuals vs Predicted by state', fontsize=16)\n",
    "    plt.show()\n",
    "    \n",
    "def plot_residuals(X_test, residuals, model_name=None):\n",
    "    plt.figure(figsize=(18, 6))\n",
    "    sns.scatterplot(x=df['date'], y=residuals, color='none', edgecolor='r')\n",
    "    if model_name:\n",
    "        plt.title(f'{model_name}: Residuals', fontsize=16)\n",
    "    else:\n",
    "        plt.title('Residuals', fontsize=16)\n",
    "    plt.show()\n",
    "    \n",
    "def plot_residuals_w_state(X_test, residuals, unpivoted_test_df, model_name=None):\n",
    "    plt.figure(figsize=(18, 6))\n",
    "    sns.scatterplot(x=df['date'], y=residuals, hue=unpivoted_test_df['state_name'])\n",
    "    if model_name:\n",
    "        plt.title(f'{model_name}: Residuals by state', fontsize=16)\n",
    "    else:\n",
    "        plt.title('Residuals by state', fontsize=16)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27616def",
   "metadata": {},
   "source": [
    "## Modeling functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4b170eb3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_full_pipeline(model, X_train, y_train, X_test, y_test, unpivoted_test_df, model_name=None):\n",
    "    if model_name:\n",
    "        print(f'{model_name}')\n",
    "    # Cross validate and report scores:\n",
    "    scores = modeler.cv_model(model, X_train, y_train)\n",
    "    print('Cross-Validation Scores:')\n",
    "    for k, v in scores.items():\n",
    "        print(f'\\t{k}: {round(v[0],2)} +/- {round(v[1],2)}')\n",
    "    # Test data\n",
    "    start = time.time()\n",
    "    y_pred, rmse, r2, fit_time, fit_predict_time = modeler.test_model(model, X_train, y_train, X_test, y_test)\n",
    "    residuals = (y_test - y_pred)\n",
    "    print(f'Time to train model: {fit_time}')\n",
    "    # Plots\n",
    "    plot_predictions(y_test, y_pred, model_name)\n",
    "    plot_residuals(X_test, residuals, model_name)\n",
    "    plot_predictions_w_state(y_test, y_pred, unpivoted_test_df, model_name)\n",
    "    plot_residuals_w_state(X_test, residuals, unpivoted_test_df, model_name)\n",
    "    return y_pred, residuals, rmse, r2, fit_time"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7fc2bd2b",
   "metadata": {},
   "source": [
    "# Gradient Boosting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "a827824f",
   "metadata": {},
   "outputs": [],
   "source": [
    "gb_model = GradientBoostingRegressor()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8e8dee00",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Cases\n",
    "gb_model.fit(X_train, y_train_cases)\n",
    "\n",
    "y_pred_cases = gb_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "7db1e850",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The mean squared error (MSE) on test set: 2335.6488\n"
     ]
    }
   ],
   "source": [
    "mse = metrics.mean_squared_error(y_test_cases, y_pred_cases)\n",
    "print(\"The mean squared error (MSE) on test set: {:.4f}\".format(mse))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "ade488f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Adapted from https://scikit-learn.org/stable/auto_examples/ensemble/plot_gradient_boosting_regression.html#plot-training-deviance\n",
    "number_of_estimators = 100\n",
    "test_score = np.zeros((number_of_estimators,), dtype=np.float64)\n",
    "for i, y_pred_cases in enumerate(gb_model.staged_predict(X_test)):\n",
    "    test_score[i] = gb_model.loss_(y_test_cases, y_pred_cases)\n",
    "min_i = np.argmin(test_score)\n",
    "\n",
    "fig = plt.figure(figsize=(6, 6))\n",
    "plt.subplot(1, 1, 1)\n",
    "plt.title(\"Deviance\")\n",
    "plt.plot(\n",
    "    np.arange(number_of_estimators) + 1,\n",
    "    gb_model.train_score_,\n",
    "    \"b-\",\n",
    "    label=\"Training Set Deviance\",\n",
    ")\n",
    "plt.plot(\n",
    "    np.arange(number_of_estimators) + 1, test_score, \"r-\", label=\"Test Set Deviance\"\n",
    ")\n",
    "\n",
    "plt.axvline(x=(min_i + 1), color='gray')\n",
    "\n",
    "plt.legend(loc=\"upper right\")\n",
    "plt.xlabel(\"Boosting Iterations\")\n",
    "plt.ylabel(\"Deviance\")\n",
    "fig.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "9cb7d133",
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_gb_depth(depths):\n",
    "    result_dict = {}\n",
    "    for d in depths:\n",
    "        print(f'Running GB w max depth {d}...')\n",
    "        gb_model = GradientBoostingRegressor(random_state=1, max_depth=d)\n",
    "        # Test various values for tree depth (default is 3)\n",
    "        # Cases\n",
    "        start = time.time()\n",
    "        gb_model.fit(X_train, y_train_cases)\n",
    "        fit_time = time.time() - start\n",
    "        print(f'Fit time for max depth {d} was {fit_time:.2f} seconds')\n",
    "        y_pred_cases = gb_model.predict(X_test)\n",
    "        # Adapted from https://scikit-learn.org/stable/auto_examples/ensemble/plot_gradient_boosting_regression.html#plot-training-deviance\n",
    "        number_of_estimators = 100\n",
    "        test_score = np.zeros((number_of_estimators,), dtype=np.float64)\n",
    "        for i, y_pred_cases in enumerate(gb_model.staged_predict(X_test)):\n",
    "            test_score[i] = gb_model.loss_(y_test_cases, y_pred_cases)\n",
    "        min_i = np.argmin(test_score)\n",
    "        \n",
    "        result_dict[d] = {\n",
    "            'test_score' : test_score\n",
    "            , 'train_score' : gb_model.train_score_\n",
    "            , 'fit_time' : fit_time\n",
    "        }\n",
    "    return result_dict\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "64ab81ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "depths = [2,3,4,5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "65b120b0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running GB w max depth 2...\n",
      "Fit time for max depth 2 was 195.59 seconds\n",
      "Running GB w max depth 3...\n",
      "Fit time for max depth 3 was 290.59 seconds\n",
      "Running GB w max depth 4...\n",
      "Fit time for max depth 4 was 380.85 seconds\n",
      "Running GB w max depth 5...\n",
      "Fit time for max depth 5 was 472.78 seconds\n"
     ]
    }
   ],
   "source": [
    "depth_results = test_gb_depth(depths)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "0d36fe61",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting max_depth results\n",
    "fig, axs = plt.subplots(nrows=1, ncols=4, sharey=True, figsize=(16, 4))\n",
    "\n",
    "for i, d in zip(range(0, len(depths)), depths):\n",
    "    axs[i].plot(\n",
    "        np.arange(number_of_estimators) + 1,\n",
    "        depth_results[depths[i]]['train_score'],\n",
    "        \"b-\",\n",
    "        label=\"Training Set Deviance\",\n",
    "    )\n",
    "    axs[i].plot(\n",
    "        np.arange(number_of_estimators) + 1, depth_results[depths[i]]['test_score'], \"r-\", label=\"Test Set Deviance\"\n",
    "    )\n",
    "    \n",
    "    axs[i].axvline(x=(np.argmin(depth_results[depths[i]]['test_score']) + 1), color='gray')\n",
    "    axs[i].set_xlabel('Boosting Iterations')\n",
    "    axs[i].set_title(f'Max Depth of {depths[i]}', fontsize=16);\n",
    "axs[0].set_ylabel('Deviance')\n",
    "fig.suptitle('GB Model performance wrt Max Tree Depth (100 boosting iterations)', fontsize=20)\n",
    "fig.tight_layout()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "d3769f54",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_times = [x[1]['fit_time'] for x in depth_results.items()]\n",
    "sns.scatterplot(x=depths, y=train_times);\n",
    "plt.title('Train time for GB models', fontsize=16);\n",
    "plt.ylabel('Train Time (sec)')\n",
    "plt.xlabel('Max Depth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "e0498a82",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'test_score': array([10877.1072729 ,  9692.55896658,  8571.00780068,  7709.15326897,\n",
      "        6857.47911935,  6163.85052313,  5596.26221274,  5057.16150171,\n",
      "        4655.26291532,  4248.52123611,  3928.04315176,  3655.80855426,\n",
      "        3413.02433172,  3167.07539318,  2971.53316314,  2816.83563247,\n",
      "        2658.58130855,  2527.90813763,  2402.25279122,  2286.74519579,\n",
      "        2201.15720328,  2104.64627996,  2034.08441655,  1975.58082236,\n",
      "        1963.23597853,  1870.88564523,  1811.41208804,  1750.40274461,\n",
      "        1748.15188831,  1780.79096702,  1731.30445425,  1695.52515208,\n",
      "        1698.67050415,  1737.13433312,  1709.67792563,  1715.98943821,\n",
      "        1694.01144226,  1732.01100249,  1719.58171997,  1704.42284136,\n",
      "        1696.55876799,  1709.39853046,  1702.91040564,  1712.29336662,\n",
      "        1698.2435618 ,  1707.65744943,  1701.68832303,  1746.87392021,\n",
      "        1735.04630471,  1769.55838352,  1749.49451572,  1739.4750038 ,\n",
      "        1748.05485599,  1739.03212253,  1737.4694837 ,  1746.4358434 ,\n",
      "        1740.61985973,  1744.98428279,  1729.84415091,  1726.93603801,\n",
      "        1719.66253717,  1711.33976293,  1707.92459833,  1707.90365452,\n",
      "        1700.72180167,  1704.94484402,  1700.68335561,  1694.44104818,\n",
      "        1728.28894249,  1734.39310261,  1730.66208346,  1721.95923268,\n",
      "        1717.13252677,  1718.14918705,  1718.13186321,  1709.2807902 ,\n",
      "        1734.31794271,  1740.49298389,  1739.78967677,  1739.78450878,\n",
      "        1731.29375643,  1726.67463684,  1733.06004933,  1754.83988102,\n",
      "        1754.82337248,  1761.81201938,  1763.96146184,  1766.44910804,\n",
      "        1789.51391263,  1789.51292468,  1781.77480594,  1782.83369291,\n",
      "        1789.17899319,  1860.89227004,  1856.95365471,  1852.91401988,\n",
      "        1851.05757109,  1851.0275927 ,  1855.81690673,  1853.01697478]), 'train_score': array([6197.47105663, 5271.85903963, 4490.67316253, 3848.29907074,\n",
      "       3301.47884457, 2858.47668679, 2482.02063107, 2162.42322425,\n",
      "       1901.52366967, 1676.98828615, 1492.24464819, 1335.71229003,\n",
      "       1206.83193727, 1092.56641549, 1000.85999687,  922.68697018,\n",
      "        855.02406548,  800.39446518,  753.15722519,  713.05288186,\n",
      "        679.54222905,  648.71745207,  624.01113844,  602.67650678,\n",
      "        583.40612485,  558.39934478,  542.5951086 ,  524.48708359,\n",
      "        511.74898978,  501.19924157,  488.01633363,  477.13804945,\n",
      "        468.65167578,  461.37957093,  453.50509847,  446.94278315,\n",
      "        441.79881853,  436.80824534,  432.18787479,  427.05431586,\n",
      "        423.28748656,  418.97847937,  415.81875991,  411.24540111,\n",
      "        406.71117423,  402.7883097 ,  400.41232585,  397.49855661,\n",
      "        394.19994276,  391.14648508,  388.85845789,  386.20642685,\n",
      "        383.31085139,  381.46313781,  378.88879839,  376.12610433,\n",
      "        374.03041545,  372.20148373,  370.63012365,  368.02277175,\n",
      "        366.27689795,  364.50965193,  363.38131355,  362.4736252 ,\n",
      "        361.24082342,  360.09866116,  358.76807717,  357.6958063 ,\n",
      "        355.59783061,  354.23870449,  353.14168887,  351.26125506,\n",
      "        350.39150836,  348.62052461,  348.05319676,  346.80810249,\n",
      "        345.1509021 ,  343.9491727 ,  342.48538281,  341.9973554 ,\n",
      "        341.24963206,  340.22642771,  339.0936132 ,  337.80312394,\n",
      "        337.39139229,  336.19406803,  334.96125312,  334.30342968,\n",
      "        333.14982855,  332.79431314,  331.84821706,  330.66608166,\n",
      "        329.87891617,  328.33287452,  327.54500336,  326.9863657 ,\n",
      "        326.28569005,  325.98966805,  324.91094748,  324.22832672]), 'fit_time': 195.58510375022888}\n",
      "{'test_score': array([10646.38835304,  9295.69951507,  8136.35023461,  7183.70801853,\n",
      "        6362.59997845,  5686.50332252,  5110.037761  ,  4606.63590534,\n",
      "        4192.94818222,  3825.33891987,  3512.16128754,  3292.62442068,\n",
      "        3105.01463938,  2912.05597571,  2749.38729474,  2663.38670832,\n",
      "        2553.60035834,  2465.56576703,  2375.21402483,  2269.23255288,\n",
      "        2198.75963476,  2171.14581063,  2180.64687254,  2161.91307101,\n",
      "        2149.27238599,  2128.22496836,  2093.04129567,  2124.95110136,\n",
      "        2111.7046937 ,  2096.38336249,  2121.60463783,  2104.64739861,\n",
      "        2066.02462841,  2026.13069105,  2094.34168017,  2073.62562372,\n",
      "        2058.63292333,  2058.96141648,  2034.79609342,  2066.5630496 ,\n",
      "        2044.94232697,  2006.56177224,  2078.38448833,  2065.58918218,\n",
      "        2107.43255771,  2101.30640661,  2112.62736183,  2096.09528994,\n",
      "        2090.52708413,  2116.65485674,  2110.90134597,  2111.84126108,\n",
      "        2107.89403869,  2120.53393671,  2120.97100895,  2139.9420366 ,\n",
      "        2123.8066276 ,  2136.50599184,  2134.56866964,  2148.82858804,\n",
      "        2144.62707166,  2151.52320653,  2168.94103155,  2192.56254149,\n",
      "        2189.91635434,  2192.26867702,  2190.91505224,  2201.76517462,\n",
      "        2205.53332447,  2205.54939608,  2207.48060137,  2285.21906772,\n",
      "        2276.17276199,  2267.15619986,  2267.84395343,  2264.82649644,\n",
      "        2255.68025349,  2270.77116092,  2283.27577388,  2298.46717509,\n",
      "        2296.28492632,  2308.13677815,  2296.07653081,  2292.82533173,\n",
      "        2286.07531323,  2285.35579615,  2281.22785435,  2284.78294687,\n",
      "        2283.20008316,  2295.35735746,  2292.00005061,  2362.26116003,\n",
      "        2364.58876632,  2360.56136026,  2374.10777971,  2366.20004324,\n",
      "        2370.1688483 ,  2358.15960448,  2356.6258665 ,  2349.87549598]), 'train_score': array([6073.60964146, 5052.02500129, 4214.10327584, 3528.40935623,\n",
      "       2970.87605162, 2515.23485293, 2143.77407621, 1839.47001893,\n",
      "       1588.19939815, 1385.89893654, 1219.23649524, 1082.73245693,\n",
      "        971.3255416 ,  871.52081418,  793.24263048,  726.7779006 ,\n",
      "        672.06567105,  626.62234449,  586.27681383,  554.20261572,\n",
      "        526.80971659,  500.02391395,  481.36652529,  461.82695652,\n",
      "        445.45774145,  430.87018708,  416.7000837 ,  406.01220203,\n",
      "        396.22618088,  386.67845218,  378.39950065,  371.88165559,\n",
      "        364.75332041,  358.08677545,  353.63786053,  348.04063238,\n",
      "        343.2689369 ,  339.14799469,  335.21350111,  331.71653705,\n",
      "        328.57091768,  324.83963741,  322.27246971,  319.86741859,\n",
      "        317.57264065,  315.30520223,  312.4486191 ,  310.55877362,\n",
      "        309.28902324,  307.24331557,  305.95158568,  303.11683184,\n",
      "        301.59262335,  299.3300836 ,  297.04517434,  295.29433764,\n",
      "        294.29406305,  292.40867888,  290.8457994 ,  289.26997581,\n",
      "        287.61963748,  286.03712729,  284.79573842,  283.56264907,\n",
      "        281.29763932,  280.18433857,  279.14554758,  278.095703  ,\n",
      "        276.64250899,  275.36080112,  274.5909592 ,  273.25098565,\n",
      "        271.96444904,  271.14296988,  270.40699138,  269.50893632,\n",
      "        268.3523745 ,  267.55120867,  266.82867706,  265.41703729,\n",
      "        264.45168928,  263.80241292,  262.25251673,  261.58388129,\n",
      "        260.44426998,  259.75444404,  259.08213962,  258.46091372,\n",
      "        257.96419362,  257.38721106,  256.56578596,  255.55331691,\n",
      "        255.04058881,  254.40928335,  253.91586779,  253.35451674,\n",
      "        252.80727325,  252.06915743,  251.5479341 ,  250.69595034]), 'fit_time': 290.5943760871887}\n",
      "{'test_score': array([10614.27925683,  9232.77918979,  8059.88592898,  7080.15774781,\n",
      "        6138.91680647,  5427.27507416,  4936.72508839,  4467.64450511,\n",
      "        4092.87015998,  3762.16046952,  3528.14161076,  3305.42371052,\n",
      "        3190.27880479,  2961.70186132,  2791.61866545,  2697.33002125,\n",
      "        2624.30483473,  2493.53401686,  2458.12315255,  2403.75839283,\n",
      "        2358.44699785,  2279.02617909,  2233.61373106,  2187.72561595,\n",
      "        2146.95596607,  2085.58294205,  2068.70362527,  2105.29445045,\n",
      "        2085.06752849,  2058.19719282,  2101.52856431,  2102.40468147,\n",
      "        2139.89788321,  2129.23117649,  2121.92529963,  2087.59074852,\n",
      "        2118.84333647,  2098.50838032,  2075.05525649,  2064.39813722,\n",
      "        2059.36576415,  2049.80559476,  2091.28243902,  2065.86444422,\n",
      "        2025.65217028,  2001.80212267,  1998.99987144,  1987.79215869,\n",
      "        1999.11207146,  1983.32404581,  1961.78081616,  1957.2472276 ,\n",
      "        1957.56404195,  1938.06723763,  1948.9215932 ,  1957.06697112,\n",
      "        1992.79011008,  2003.31839524,  1993.8269952 ,  2017.95044149,\n",
      "        2025.60942403,  2009.20519888,  2002.9067817 ,  2001.49995164,\n",
      "        1995.39341492,  2009.42278858,  2011.46317962,  2016.90554944,\n",
      "        2015.69237025,  2053.20621547,  2055.56460837,  2060.69191577,\n",
      "        2065.22096837,  2085.04805811,  2086.20115087,  2094.23645831,\n",
      "        2096.45853709,  2095.52714948,  2086.39596328,  2108.61948613,\n",
      "        2115.94802447,  2114.06017765,  2112.55905119,  2105.85311748,\n",
      "        2100.8252845 ,  2133.11400171,  2128.79279701,  2129.57321427,\n",
      "        2139.43767711,  2175.29814284,  2180.54084122,  2173.09702296,\n",
      "        2167.84300304,  2165.71666413,  2162.65594481,  2154.83487513,\n",
      "        2192.69227726,  2200.44579385,  2200.83567086,  2199.5304257 ]), 'train_score': array([6033.35122537, 4984.42378867, 4131.20480596, 3440.337554  ,\n",
      "       2872.84034377, 2411.48518955, 2037.37285034, 1732.11045443,\n",
      "       1483.13988409, 1277.99911145, 1111.24844423,  975.901449  ,\n",
      "        864.62688884,  766.76404049,  687.37762038,  624.17431962,\n",
      "        568.16407092,  522.71120152,  485.36585034,  454.76454557,\n",
      "        427.97992524,  406.47131805,  387.23835458,  372.0115255 ,\n",
      "        356.87997101,  345.61003585,  334.98067165,  324.81482601,\n",
      "        316.43355333,  308.75538914,  301.19553693,  294.7575938 ,\n",
      "        290.27246154,  285.91417645,  281.75987928,  275.96194256,\n",
      "        272.95474904,  268.28637257,  264.34022552,  262.40998174,\n",
      "        258.76334388,  255.62765244,  253.77671942,  251.99040928,\n",
      "        249.52957214,  247.9327624 ,  246.052436  ,  244.65762476,\n",
      "        242.66287519,  240.38675493,  239.33651374,  238.01874385,\n",
      "        235.66391473,  234.71475691,  233.18097227,  231.260141  ,\n",
      "        230.23574179,  228.92049512,  228.17213041,  226.96754994,\n",
      "        226.12281593,  223.9542135 ,  223.32857752,  221.7077    ,\n",
      "        220.73134521,  219.66915268,  218.62172986,  217.85704766,\n",
      "        216.90230511,  216.12399798,  215.00875163,  213.8562731 ,\n",
      "        213.25801728,  212.50123677,  211.76361483,  210.78232595,\n",
      "        210.07508281,  209.54707678,  208.9918206 ,  208.30507126,\n",
      "        207.77508308,  206.8262213 ,  206.19980676,  205.56224036,\n",
      "        204.76815163,  204.14314043,  203.46355468,  202.94203557,\n",
      "        202.62585551,  202.20439358,  201.24331434,  200.62399266,\n",
      "        200.10542403,  199.60025797,  199.16044729,  198.4807852 ,\n",
      "        197.93557671,  197.35187971,  196.88294727,  196.48449133]), 'fit_time': 380.8475089073181}\n",
      "{'test_score': array([10753.05592389,  9565.74544184,  8409.1869982 ,  7492.06765278,\n",
      "        6635.89265476,  5982.023356  ,  5406.02542949,  4891.61704257,\n",
      "        4488.95262624,  4203.51880891,  3894.60518722,  3625.69216411,\n",
      "        3446.80791553,  3297.73933163,  3173.44994654,  2996.18963258,\n",
      "        2932.96367081,  2833.03559666,  2716.07609222,  2630.4409796 ,\n",
      "        2598.48452179,  2547.86076549,  2512.55282158,  2466.30323842,\n",
      "        2477.41609483,  2476.94025333,  2397.0833081 ,  2350.87911723,\n",
      "        2334.72030403,  2312.67965578,  2280.0791261 ,  2213.49565847,\n",
      "        2230.05392169,  2268.21250293,  2275.6366413 ,  2259.5472162 ,\n",
      "        2236.62882559,  2221.062887  ,  2210.45149335,  2194.89282571,\n",
      "        2201.50211273,  2208.36151356,  2184.37043905,  2180.02120018,\n",
      "        2164.81690072,  2171.89545483,  2171.56981227,  2167.9953755 ,\n",
      "        2176.00555127,  2171.05335921,  2157.64070672,  2162.27608069,\n",
      "        2146.63515194,  2145.37560255,  2152.99147084,  2147.5296524 ,\n",
      "        2142.78402447,  2146.80997807,  2150.44816862,  2147.88348465,\n",
      "        2141.80653578,  2132.6925386 ,  2134.77914571,  2130.36644565,\n",
      "        2147.23357473,  2146.172867  ,  2150.24563158,  2151.09681319,\n",
      "        2149.46688687,  2142.89174735,  2143.48657615,  2142.80600071,\n",
      "        2164.1914336 ,  2161.5326893 ,  2168.35838876,  2172.14399746,\n",
      "        2173.97974113,  2203.65452653,  2205.67989406,  2211.13708942,\n",
      "        2210.51772197,  2214.26370101,  2212.44024691,  2211.82262531,\n",
      "        2209.04453028,  2220.40375236,  2222.74107259,  2224.44434822,\n",
      "        2223.94982813,  2223.848717  ,  2222.79848365,  2224.13074815,\n",
      "        2223.77507597,  2233.11810181,  2230.34545013,  2226.00369034,\n",
      "        2244.5694846 ,  2247.53483537,  2247.3835043 ,  2247.28330062]), 'train_score': array([6017.25549767, 4954.85170836, 4088.81354223, 3387.13905892,\n",
      "       2814.79349592, 2349.78672599, 1971.77886455, 1663.61354793,\n",
      "       1413.76917819, 1208.85459287, 1041.82253066,  904.91031419,\n",
      "        791.13582336,  699.39218138,  623.94663435,  561.10331404,\n",
      "        504.11352146,  461.90683952,  423.45815151,  394.41275641,\n",
      "        365.89033322,  345.36023578,  325.23807508,  310.90872865,\n",
      "        299.46029573,  285.94537114,  276.26386069,  266.90617284,\n",
      "        259.5978771 ,  252.44964643,  245.65717399,  241.10798195,\n",
      "        235.52610484,  231.51972787,  226.46076872,  223.77168692,\n",
      "        220.87243752,  218.66786832,  216.60066257,  213.42134474,\n",
      "        210.53816812,  208.38181532,  205.1619626 ,  202.76677549,\n",
      "        200.82536834,  199.56948329,  197.35269229,  195.59290218,\n",
      "        194.61777796,  193.11775771,  192.18262351,  190.81406487,\n",
      "        190.06296534,  188.48310138,  187.00034167,  185.83220453,\n",
      "        184.0967848 ,  182.84765164,  181.25052966,  180.19541457,\n",
      "        179.17291022,  178.42358215,  177.42191473,  176.59065877,\n",
      "        175.93877404,  174.92893335,  174.09224173,  173.39300231,\n",
      "        172.6169825 ,  171.69495336,  171.07532   ,  170.55296948,\n",
      "        169.96072541,  169.35192723,  168.66038071,  168.0880497 ,\n",
      "        167.11476782,  166.56953694,  165.61615572,  164.9335496 ,\n",
      "        164.28220183,  163.83966575,  163.24248208,  162.63058393,\n",
      "        161.84862061,  161.27252122,  160.51917517,  160.1708128 ,\n",
      "        159.96609131,  158.7371356 ,  158.14771021,  157.76534122,\n",
      "        157.12444152,  156.71894514,  156.15422067,  155.39086487,\n",
      "        154.98915893,  154.63309379,  154.30314732,  153.81559749]), 'fit_time': 472.77875113487244}\n"
     ]
    }
   ],
   "source": [
    "for x in depth_results.items():\n",
    "    print(x[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae454650",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}