final pipeline.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 245,
   "id": "c1a5fd80",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "id": "145dc17c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, cross_val_score\n",
    "from sklearn.ensemble import RandomForestRegressor, ExtraTreesRegressor, HistGradientBoostingRegressor\n",
    "from sklearn.ensemble import AdaBoostRegressor, GradientBoostingRegressor\n",
    "from sklearn.linear_model import LinearRegression, Ridge, Lasso, ElasticNet\n",
    "from sklearn.metrics import mean_absolute_error\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.feature_selection import SelectFromModel,RFECV\n",
    "from catboost import CatBoostRegressor\n",
    "from xgboost import XGBRegressor\n",
    "from lightgbm import LGBMRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "7a5753e2",
   "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>mean</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>oem</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>audi</th>\n",
       "      <td>2.021626e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bmw</th>\n",
       "      <td>2.642810e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>chevrolet</th>\n",
       "      <td>1.917756e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>datsun</th>\n",
       "      <td>2.861576e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fiat</th>\n",
       "      <td>2.361982e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>force</th>\n",
       "      <td>5.703524e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ford</th>\n",
       "      <td>5.973613e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>honda</th>\n",
       "      <td>5.519711e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hyundai</th>\n",
       "      <td>5.507323e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isuzu</th>\n",
       "      <td>1.325532e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jaguar</th>\n",
       "      <td>2.991708e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jeep</th>\n",
       "      <td>1.695172e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kia</th>\n",
       "      <td>1.424329e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>land rover</th>\n",
       "      <td>4.190503e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>lexus</th>\n",
       "      <td>7.163349e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra</th>\n",
       "      <td>8.781328e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra renault</th>\n",
       "      <td>1.166455e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra ssangyong</th>\n",
       "      <td>6.722756e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>maruti</th>\n",
       "      <td>4.707731e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mercedes-benz</th>\n",
       "      <td>2.992328e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mg</th>\n",
       "      <td>1.769030e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mini</th>\n",
       "      <td>2.582763e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mitsubishi</th>\n",
       "      <td>8.504703e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nissan</th>\n",
       "      <td>4.538965e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>porsche</th>\n",
       "      <td>4.188174e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>renault</th>\n",
       "      <td>4.923059e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>skoda</th>\n",
       "      <td>8.065404e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tata</th>\n",
       "      <td>7.481152e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>toyota</th>\n",
       "      <td>1.173012e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volkswagen</th>\n",
       "      <td>5.370024e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volvo</th>\n",
       "      <td>2.562063e+06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            mean\n",
       "oem                             \n",
       "audi                2.021626e+06\n",
       "bmw                 2.642810e+06\n",
       "chevrolet           1.917756e+05\n",
       "datsun              2.861576e+05\n",
       "fiat                2.361982e+05\n",
       "force               5.703524e+05\n",
       "ford                5.973613e+05\n",
       "honda               5.519711e+05\n",
       "hyundai             5.507323e+05\n",
       "isuzu               1.325532e+06\n",
       "jaguar              2.991708e+06\n",
       "jeep                1.695172e+06\n",
       "kia                 1.424329e+06\n",
       "land rover          4.190503e+06\n",
       "lexus               7.163349e+06\n",
       "mahindra            8.781328e+05\n",
       "mahindra renault    1.166455e+05\n",
       "mahindra ssangyong  6.722756e+05\n",
       "maruti              4.707731e+05\n",
       "mercedes-benz       2.992328e+06\n",
       "mg                  1.769030e+06\n",
       "mini                2.582763e+06\n",
       "mitsubishi          8.504703e+05\n",
       "nissan              4.538965e+05\n",
       "porsche             4.188174e+06\n",
       "renault             4.923059e+05\n",
       "skoda               8.065404e+05\n",
       "tata                7.481152e+05\n",
       "toyota              1.173012e+06\n",
       "volkswagen          5.370024e+05\n",
       "volvo               2.562063e+06"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.groupby('oem')['listed_price'].agg([np.mean])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "id": "9b8843a5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['myear', 'transmission', 'fuel', 'km', 'oem', 'model', 'variant',\n",
       "       'Valve Configuration', 'Turbo Charger', 'Kerb Weight', 'Drive Type',\n",
       "       'Seats', 'Top Speed', 'Acceleration', 'Doors', 'Cargo Volume',\n",
       "       'owner_type', 'Fuel Supply System', 'Max Torque Delivered', 'body',\n",
       "       'Gear Box', 'Steering Type', 'Front Brake Type', 'Rear Brake Type',\n",
       "       'Tyre Type', 'exterior_color', 'listed_price', 'avg_measure',\n",
       "       'Features', 'Valves', 'Tread'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "id": "facadb2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_csv('final_train2.csv')\n",
    "test = pd.read_csv('final_test2.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "id": "a053f4d8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat([train, test], axis = 0)\n",
    "X = df.drop('listed_price', axis = 1)\n",
    "y = df['listed_price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "id": "26c1bb10",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train_test_split\n",
    "X_train = train.drop('listed_price', axis = 1)\n",
    "X_test = test.drop('listed_price', axis = 1)\n",
    "y_train = train['listed_price']\n",
    "y_test = test['listed_price']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "id": "a87bcaf7",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestRegressor()\n",
    "et = ExtraTreesRegressor()\n",
    "hgb = HistGradientBoostingRegressor()\n",
    "cat = CatBoostRegressor(verbose = False)\n",
    "xg = XGBRegressor()\n",
    "lg = LGBMRegressor()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "id": "372c526a",
   "metadata": {},
   "outputs": [],
   "source": [
    "le = LabelEncoder()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "id": "83a38a46",
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (707217535.py, line 2)",
     "output_type": "error",
     "traceback": [
      "\u001b[1;36m  File \u001b[1;32m\"C:\\Users\\rajni\\AppData\\Local\\Temp\\ipykernel_24484\\707217535.py\"\u001b[1;36m, line \u001b[1;32m2\u001b[0m\n\u001b[1;33m    temp_test = pd.DataFrame(le.transform(X_test[cat_cols])\u001b[0m\n\u001b[1;37m    ^\u001b[0m\n\u001b[1;31mSyntaxError\u001b[0m\u001b[1;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": [
    "temp_train = pd.DataFrame(le.fit_transform(X_train[cat_cols])\n",
    "temp_test = pd.DataFrame(le.transform(X_test[cat_cols])\n",
    "temp = pd.concat([temp_train, temp_test], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "id": "2048d828",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(33760, 64)"
      ]
     },
     "execution_count": 290,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "temp_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "id": "987be81f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3774, 57)"
      ]
     },
     "execution_count": 291,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "temp_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c4086b5",
   "metadata": {},
   "source": [
    "## 1. Predictions and evaluations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "id": "3928cf36",
   "metadata": {},
   "outputs": [],
   "source": [
    "num_cols = [col for col in X_train.columns if X_train[col].dtype != 'O']\n",
    "cat_cols = [col for col in X_train.columns if X_train[col].dtype == 'O' and col not in nope]\n",
    "nope = ['oem', 'variant', 'model']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "id": "c5a986c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import RobustScaler\n",
    "rs = RobustScaler()\n",
    "rX_train = pd.DataFrame(rs.fit_transform(X_train[num_cols]))\n",
    "rX_test = pd.DataFrame(rs.transform(X_test[num_cols]))\n",
    "rX_train.columns = num_cols\n",
    "rX_test.columns = num_cols\n",
    "rX = pd.concat([rX_train, rX_test], axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "id": "0368a8f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, X, X_train, X_test):\n",
    "    model.fit(X_train, y_train)\n",
    "    testpreds = model.predict(X_test)\n",
    "    trainpreds = model.predict(X_train)\n",
    "    scoretest = mean_absolute_error(y_test, testpreds)\n",
    "    scoretrain = mean_absolute_error(y_train, trainpreds)\n",
    "    score_percenttest = scoretest * 100 / y_test.mean()\n",
    "    score_percenttrain =  scoretrain * 100 / y_train.mean()\n",
    "    \n",
    "    print(f\"for {model}\")\n",
    "    print(\"For train\")\n",
    "    print(\"MAE: \", scoretrain)\n",
    "    print(\"MAE percent: \", score_percenttrain)\n",
    "    print(\"=================================\")\n",
    "    print(\"For test\")\n",
    "    print(\"MAE: \", scoretest)\n",
    "    print(\"MAE percent: \", score_percenttest)\n",
    "    print(\"==================================\")\n",
    "    \n",
    "    arr = cross_val_score(estimator = model,\n",
    "                    X = X,\n",
    "                    y = y,\n",
    "                    cv = 5,\n",
    "                    n_jobs = -1,\n",
    "                   scoring = 'neg_mean_absolute_error')\n",
    "    values = []\n",
    "    percentages = []\n",
    "    for i,val in enumerate(arr):\n",
    "        value = -1 * val\n",
    "        percent =  value * 100 / y.mean()\n",
    "        values.append(value)\n",
    "        percentages.append(percent)\n",
    "        print(f\"for fold {i + 1}\")\n",
    "        print(\"MAE: \", value)\n",
    "        print(\"MAE percent: \", percent)\n",
    "        print(\"============================\")\n",
    "    print(\"CV Summary: \")\n",
    "    print(\"MAE: \", np.mean(values), ' ± ', np.std(values))\n",
    "    print(\"MAE percent: \", np.mean(percentages), ' ± ', np.std(percentages))\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "id": "efe73ecc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 : 2\n",
      "1 : 3\n",
      "2 : 5\n",
      "3 : 6\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4.0"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = [2, 3, 5, 6]\n",
    "for i, val in enumerate(a):\n",
    "    print(f'{i} : {val}')\n",
    "np.mean(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "ce6a48c9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for RandomForestRegressor()\n",
      "For train\n",
      "MAE:  37242.735578312015\n",
      "MAE percent:  4.822322795897742\n",
      "=================================\n",
      "For test\n",
      "MAE:  102752.18512380133\n",
      "MAE percent:  12.98342647695737\n",
      "==================================\n",
      "for fold 1\n",
      "MAE:  100226.837675376\n",
      "MAE percent:  12.945518464902444\n",
      "============================\n",
      "for fold 2\n",
      "MAE:  102269.78237244285\n",
      "MAE percent:  13.209389688539293\n",
      "============================\n",
      "for fold 3\n",
      "MAE:  97468.32767955624\n",
      "MAE percent:  12.589223255807205\n",
      "============================\n",
      "for fold 4\n",
      "MAE:  102252.27498692332\n",
      "MAE percent:  13.20712839617717\n",
      "============================\n",
      "for fold 5\n",
      "MAE:  100901.91648835852\n",
      "MAE percent:  13.032713126945312\n",
      "============================\n",
      "CV Summary: \n",
      "MAE:  100623.82784053139  ±  1763.6962427190517\n",
      "MAE percent:  12.996794586474286  ±  0.2278028800085345\n",
      "Wall time: 43.8 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(rf, X[num_cols], X_train[num_cols], X_test[num_cols])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "6d239a77",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for ExtraTreesRegressor()\n",
      "For train\n",
      "MAE:  574.2409952606636\n",
      "MAE percent:  0.07435478083938363\n",
      "=================================\n",
      "For test\n",
      "MAE:  101847.32625022081\n",
      "MAE percent:  12.869091500596525\n",
      "==================================\n",
      "for fold 1\n",
      "MAE:  103303.44426601838\n",
      "MAE percent:  13.342899728765135\n",
      "============================\n",
      "for fold 2\n",
      "MAE:  104766.94920674036\n",
      "MAE percent:  13.531929240949845\n",
      "============================\n",
      "for fold 3\n",
      "MAE:  101407.57341081658\n",
      "MAE percent:  13.098024885536242\n",
      "============================\n",
      "for fold 4\n",
      "MAE:  106100.54781959059\n",
      "MAE percent:  13.704179766535988\n",
      "============================\n",
      "for fold 5\n",
      "MAE:  104523.53491162625\n",
      "MAE percent:  13.500489316024472\n",
      "============================\n",
      "CV Summary: \n",
      "MAE:  104020.40992295844  ±  1579.7711048041085\n",
      "MAE percent:  13.435504587562338  ±  0.2040467052726868\n",
      "Wall time: 28 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(et, X[num_cols], X_train[num_cols], X_test[num_cols])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "id": "ae187883",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for HistGradientBoostingRegressor()\n",
      "For train\n",
      "MAE:  97118.78651630798\n",
      "MAE percent:  12.575288330867158\n",
      "=================================\n",
      "For test\n",
      "MAE:  112608.08616082298\n",
      "MAE percent:  14.228785554471502\n",
      "==================================\n",
      "for fold 1\n",
      "MAE:  106486.24456949692\n",
      "MAE percent:  13.753997205792501\n",
      "============================\n",
      "for fold 2\n",
      "MAE:  105563.52295354895\n",
      "MAE percent:  13.634816455463852\n",
      "============================\n",
      "for fold 3\n",
      "MAE:  106222.75741356896\n",
      "MAE percent:  13.719964626081886\n",
      "============================\n",
      "for fold 4\n",
      "MAE:  111066.56144202055\n",
      "MAE percent:  14.345600992000021\n",
      "============================\n",
      "for fold 5\n",
      "MAE:  108903.19177400338\n",
      "MAE percent:  14.066175414646832\n",
      "============================\n",
      "CV Summary: \n",
      "MAE:  107648.45563052777  ±  2048.8096227839633\n",
      "MAE percent:  13.904110938797018  ±  0.26462875032258776\n",
      "Wall time: 3.61 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(hgb, X[num_cols], X_train[num_cols], X_test[num_cols])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "id": "fcd0de0f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for <catboost.core.CatBoostRegressor object at 0x00000259E9B167C0>\n",
      "For train\n",
      "MAE:  82131.45310054386\n",
      "MAE percent:  10.634674719694962\n",
      "=================================\n",
      "For test\n",
      "MAE:  97890.36360292509\n",
      "MAE percent:  12.369102779759773\n",
      "==================================\n",
      "for fold 1\n",
      "MAE:  96511.84952495171\n",
      "MAE percent:  12.465682436812067\n",
      "============================\n",
      "for fold 2\n",
      "MAE:  96738.7232086519\n",
      "MAE percent:  12.494985940041952\n",
      "============================\n",
      "for fold 3\n",
      "MAE:  94805.32822232165\n",
      "MAE percent:  12.245264397628791\n",
      "============================\n",
      "for fold 4\n",
      "MAE:  97828.27436478926\n",
      "MAE percent:  12.635714760160129\n",
      "============================\n",
      "for fold 5\n",
      "MAE:  97583.64236362594\n",
      "MAE percent:  12.60411755364721\n",
      "============================\n",
      "CV Summary: \n",
      "MAE:  96693.5635368681  ±  1065.8612444429007\n",
      "MAE percent:  12.48915301765803  ±  0.13766897909769604\n",
      "Wall time: 28.9 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(cat, X[num_cols], X_train[num_cols], X_test[num_cols])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "id": "42aa5f2c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for XGBRegressor(base_score=None, booster=None, callbacks=None,\n",
      "             colsample_bylevel=None, colsample_bynode=None,\n",
      "             colsample_bytree=None, early_stopping_rounds=None,\n",
      "             enable_categorical=False, eval_metric=None, feature_types=None,\n",
      "             gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
      "             interaction_constraints=None, learning_rate=None, max_bin=None,\n",
      "             max_cat_threshold=None, max_cat_to_onehot=None,\n",
      "             max_delta_step=None, max_depth=None, max_leaves=None,\n",
      "             min_child_weight=None, missing=nan, monotone_constraints=None,\n",
      "             n_estimators=100, n_jobs=None, num_parallel_tree=None,\n",
      "             predictor=None, random_state=None, ...)\n",
      "For train\n",
      "MAE:  75787.64678443078\n",
      "MAE percent:  9.813255956118232\n",
      "=================================\n",
      "For test\n",
      "MAE:  101774.35136998212\n",
      "MAE percent:  12.859870635939464\n",
      "==================================\n",
      "for fold 1\n",
      "MAE:  98063.98895614094\n",
      "MAE percent:  12.666160174438023\n",
      "============================\n",
      "for fold 2\n",
      "MAE:  98232.78158822599\n",
      "MAE percent:  12.687961801487996\n",
      "============================\n",
      "for fold 3\n",
      "MAE:  97632.18262824489\n",
      "MAE percent:  12.610387120825923\n",
      "============================\n",
      "for fold 4\n",
      "MAE:  100091.62174938807\n",
      "MAE percent:  12.92805367895065\n",
      "============================\n",
      "for fold 5\n",
      "MAE:  98353.43416855058\n",
      "MAE percent:  12.70354555373097\n",
      "============================\n",
      "CV Summary: \n",
      "MAE:  98474.80181811008  ±  844.533586649916\n",
      "MAE percent:  12.71922166588671  ±  0.10908181275375911\n",
      "Wall time: 6.11 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(xg, rX[num_cols], rX_train[num_cols], rX_test[num_cols])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f8552f90",
   "metadata": {},
   "source": [
    "### Adding encoded columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "id": "af367a01",
   "metadata": {},
   "outputs": [],
   "source": [
    "nX = pd.concat([X[num_cols], temp], axis = 1)\n",
    "nX_train = pd.concat([X_train[num_cols], temp_train], axis = 1)\n",
    "nX_test = pd.concat([X_test[num_cols], temp_test], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "id": "823265e0",
   "metadata": {},
   "outputs": [
    {
     "ename": "CatBoostError",
     "evalue": "C:/Go_Agent/pipelines/BuildMaster/catboost.git/catboost/libs/data/model_dataset_compatibility.cpp:81: At position 20 should be feature with name Valve Configuration_dohc (found Valve Configuration_idsi).",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mCatBoostError\u001b[0m                             Traceback (most recent call last)",
      "\u001b[1;32m<timed eval>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Local\\Temp\\ipykernel_24484\\3914687306.py\u001b[0m in \u001b[0;36mevaluate\u001b[1;34m(model, X, X_train, X_test)\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mdef\u001b[0m \u001b[0mevaluate\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX_train\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX_test\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      2\u001b[0m     \u001b[0mmodel\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[1;32m----> 3\u001b[1;33m     \u001b[0mtestpreds\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_test\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      4\u001b[0m     \u001b[0mtrainpreds\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      5\u001b[0m     \u001b[0mscoretest\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmean_absolute_error\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtestpreds\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\catboost\\core.py\u001b[0m in \u001b[0;36mpredict\u001b[1;34m(self, data, prediction_type, ntree_start, ntree_end, thread_count, verbose, task_type)\u001b[0m\n\u001b[0;32m   5777\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mprediction_type\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   5778\u001b[0m             \u001b[0mprediction_type\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_get_default_prediction_type\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 5779\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_start\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'predict'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtask_type\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   5780\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   5781\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mstaged_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'RawFormulaVal'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_start\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0meval_period\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m-\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\catboost\\core.py\u001b[0m in \u001b[0;36m_predict\u001b[1;34m(self, data, prediction_type, ntree_start, ntree_end, thread_count, verbose, parent_method_name, task_type)\u001b[0m\n\u001b[0;32m   2544\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_validate_prediction_type\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mprediction_type\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2545\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 2546\u001b[1;33m         \u001b[0mpredictions\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_base_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_start\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtask_type\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   2547\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mpredictions\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;32mif\u001b[0m \u001b[0mdata_is_single_object\u001b[0m \u001b[1;32melse\u001b[0m \u001b[0mpredictions\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   2548\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\anaconda3\\lib\\site-packages\\catboost\\core.py\u001b[0m in \u001b[0;36m_base_predict\u001b[1;34m(self, pool, prediction_type, ntree_start, ntree_end, thread_count, verbose, task_type)\u001b[0m\n\u001b[0;32m   1811\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1812\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m_base_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpool\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_start\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtask_type\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 1813\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_object\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_base_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpool\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_start\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mtask_type\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   1814\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   1815\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m_base_virtual_ensembles_predict\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpool\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mprediction_type\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mntree_end\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mvirtual_ensembles_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mthread_count\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m_catboost.pyx\u001b[0m in \u001b[0;36m_catboost._CatBoost._base_predict\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;32m_catboost.pyx\u001b[0m in \u001b[0;36m_catboost._CatBoost._base_predict\u001b[1;34m()\u001b[0m\n",
      "\u001b[1;31mCatBoostError\u001b[0m: C:/Go_Agent/pipelines/BuildMaster/catboost.git/catboost/libs/data/model_dataset_compatibility.cpp:81: At position 20 should be feature with name Valve Configuration_dohc (found Valve Configuration_idsi)."
     ]
    }
   ],
   "source": [
    "%%time\n",
    "evaluate(cat, nX, nX_train, nX_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1cde0278",
   "metadata": {},
   "source": [
    "### Error Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "id": "4691a828",
   "metadata": {},
   "outputs": [],
   "source": [
    "cat_preds = pd.Series(cat.predict(X_test[num_cols]), name = 'predictions')\n",
    "cat_residuals = pd.Series(np.abs(y_test - cat_preds), name = 'residuals')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "aff243f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No artists with labels found to put in legend.  Note that artists whose label start with an underscore are ignored when legend() is called with no argument.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x25981106430>"
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (15, 6))\n",
    "sns.scatterplot(x = y_test, y = cat_residuals)\n",
    "plt.axvline(x = 1000000, color = 'r', linestyle = '--')\n",
    "plt.axhline(y = 200000, color = 'g', linestyle = '--')\n",
    "plt.axhline(y = 0, color = 'g', linestyle = '--')\n",
    "plt.yticks()\n",
    "plt.tight_layout()\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "ebd94911",
   "metadata": {},
   "outputs": [],
   "source": [
    "errors = pd.concat([y_test, cat_preds, cat_residuals, X_test['oem']], axis = 1)\n",
    "game = errors.loc[errors.residuals > 150000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "db92599f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='residuals', ylabel='Count'>"
      ]
     },
     "execution_count": 231,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (14, 3))\n",
    "sns.histplot(errors.residuals, fill = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "cc70d257",
   "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>mean</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>oem</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>audi</th>\n",
       "      <td>2.750409e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bmw</th>\n",
       "      <td>4.093437e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>chevrolet</th>\n",
       "      <td>4.489312e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>datsun</th>\n",
       "      <td>4.808099e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fiat</th>\n",
       "      <td>4.171439e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ford</th>\n",
       "      <td>7.331256e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>honda</th>\n",
       "      <td>6.326644e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hyundai</th>\n",
       "      <td>6.332960e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isuzu</th>\n",
       "      <td>6.568604e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jaguar</th>\n",
       "      <td>5.767772e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jeep</th>\n",
       "      <td>1.693672e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kia</th>\n",
       "      <td>9.737125e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>land rover</th>\n",
       "      <td>6.078487e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>lexus</th>\n",
       "      <td>1.459142e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra</th>\n",
       "      <td>1.167114e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra renault</th>\n",
       "      <td>3.280102e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra ssangyong</th>\n",
       "      <td>6.522290e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>maruti</th>\n",
       "      <td>5.379049e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mercedes-benz</th>\n",
       "      <td>3.408705e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mg</th>\n",
       "      <td>1.564595e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mini</th>\n",
       "      <td>4.761970e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mitsubishi</th>\n",
       "      <td>1.746206e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nissan</th>\n",
       "      <td>7.760715e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>porsche</th>\n",
       "      <td>1.041753e+06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>renault</th>\n",
       "      <td>6.554328e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>skoda</th>\n",
       "      <td>1.199271e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tata</th>\n",
       "      <td>8.407832e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>toyota</th>\n",
       "      <td>1.513558e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volkswagen</th>\n",
       "      <td>8.713161e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volvo</th>\n",
       "      <td>2.764102e+05</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            mean\n",
       "oem                             \n",
       "audi                2.750409e+05\n",
       "bmw                 4.093437e+05\n",
       "chevrolet           4.489312e+04\n",
       "datsun              4.808099e+04\n",
       "fiat                4.171439e+04\n",
       "ford                7.331256e+04\n",
       "honda               6.326644e+04\n",
       "hyundai             6.332960e+04\n",
       "isuzu               6.568604e+04\n",
       "jaguar              5.767772e+05\n",
       "jeep                1.693672e+05\n",
       "kia                 9.737125e+04\n",
       "land rover          6.078487e+05\n",
       "lexus               1.459142e+06\n",
       "mahindra            1.167114e+05\n",
       "mahindra renault    3.280102e+04\n",
       "mahindra ssangyong  6.522290e+04\n",
       "maruti              5.379049e+04\n",
       "mercedes-benz       3.408705e+05\n",
       "mg                  1.564595e+05\n",
       "mini                4.761970e+05\n",
       "mitsubishi          1.746206e+05\n",
       "nissan              7.760715e+04\n",
       "porsche             1.041753e+06\n",
       "renault             6.554328e+04\n",
       "skoda               1.199271e+05\n",
       "tata                8.407832e+04\n",
       "toyota              1.513558e+05\n",
       "volkswagen          8.713161e+04\n",
       "volvo               2.764102e+05"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "errors.groupby('oem')['residuals'].agg([np.mean])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "id": "20d26c7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "mean_errors = errors.loc[errors.listed_price < 2000000].groupby('oem')['residuals'].agg([np.mean, max, min])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "3f23c2cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "less_than_10l = errors.loc[errors.listed_price < 2000000]\n",
    "counts = less_than_10l.oem.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "da28d3e8",
   "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>mean</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "      <th>oem</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>audi</th>\n",
       "      <td>2.138273e+05</td>\n",
       "      <td>5.108954e+05</td>\n",
       "      <td>1.461933e+04</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bmw</th>\n",
       "      <td>2.453293e+05</td>\n",
       "      <td>1.717239e+06</td>\n",
       "      <td>1.580588e+02</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>chevrolet</th>\n",
       "      <td>4.489312e+04</td>\n",
       "      <td>2.416497e+05</td>\n",
       "      <td>4.295486e+02</td>\n",
       "      <td>93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>datsun</th>\n",
       "      <td>4.808099e+04</td>\n",
       "      <td>1.325812e+05</td>\n",
       "      <td>3.669611e+02</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>fiat</th>\n",
       "      <td>4.171439e+04</td>\n",
       "      <td>9.962768e+04</td>\n",
       "      <td>3.312697e+03</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ford</th>\n",
       "      <td>6.219472e+04</td>\n",
       "      <td>2.092850e+05</td>\n",
       "      <td>1.654592e+03</td>\n",
       "      <td>106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>honda</th>\n",
       "      <td>6.276095e+04</td>\n",
       "      <td>5.132928e+05</td>\n",
       "      <td>7.773015e+01</td>\n",
       "      <td>390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>hyundai</th>\n",
       "      <td>6.222251e+04</td>\n",
       "      <td>1.207814e+06</td>\n",
       "      <td>1.594018e+01</td>\n",
       "      <td>732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>isuzu</th>\n",
       "      <td>6.568604e+04</td>\n",
       "      <td>6.977165e+04</td>\n",
       "      <td>6.160043e+04</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jaguar</th>\n",
       "      <td>2.141860e+05</td>\n",
       "      <td>2.369526e+05</td>\n",
       "      <td>1.914194e+05</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jeep</th>\n",
       "      <td>1.171754e+05</td>\n",
       "      <td>3.292687e+05</td>\n",
       "      <td>5.813732e+02</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kia</th>\n",
       "      <td>8.948121e+04</td>\n",
       "      <td>3.148552e+05</td>\n",
       "      <td>1.030224e+03</td>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>land rover</th>\n",
       "      <td>4.867652e+04</td>\n",
       "      <td>4.867652e+04</td>\n",
       "      <td>4.867652e+04</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra</th>\n",
       "      <td>1.131091e+05</td>\n",
       "      <td>1.328702e+06</td>\n",
       "      <td>6.647849e+02</td>\n",
       "      <td>212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra renault</th>\n",
       "      <td>3.280102e+04</td>\n",
       "      <td>5.241530e+04</td>\n",
       "      <td>1.387338e+04</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mahindra ssangyong</th>\n",
       "      <td>6.522290e+04</td>\n",
       "      <td>1.844679e+05</td>\n",
       "      <td>3.891479e+02</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>maruti</th>\n",
       "      <td>5.379049e+04</td>\n",
       "      <td>3.395944e+05</td>\n",
       "      <td>6.581418e+01</td>\n",
       "      <td>985</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mercedes-benz</th>\n",
       "      <td>1.951546e+05</td>\n",
       "      <td>8.761309e+05</td>\n",
       "      <td>7.065350e+03</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mg</th>\n",
       "      <td>1.332216e+05</td>\n",
       "      <td>3.185510e+05</td>\n",
       "      <td>5.628516e+03</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mini</th>\n",
       "      <td>3.478913e+05</td>\n",
       "      <td>3.478913e+05</td>\n",
       "      <td>3.478913e+05</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mitsubishi</th>\n",
       "      <td>1.746206e+05</td>\n",
       "      <td>6.813291e+05</td>\n",
       "      <td>1.510951e+04</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nissan</th>\n",
       "      <td>7.760715e+04</td>\n",
       "      <td>2.494148e+05</td>\n",
       "      <td>1.326066e+03</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>porsche</th>\n",
       "      <td>1.541247e+06</td>\n",
       "      <td>1.541247e+06</td>\n",
       "      <td>1.541247e+06</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>renault</th>\n",
       "      <td>6.554328e+04</td>\n",
       "      <td>2.581085e+05</td>\n",
       "      <td>1.502180e+03</td>\n",
       "      <td>119</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>skoda</th>\n",
       "      <td>1.114758e+05</td>\n",
       "      <td>6.627035e+05</td>\n",
       "      <td>2.155874e+03</td>\n",
       "      <td>81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tata</th>\n",
       "      <td>7.780391e+04</td>\n",
       "      <td>3.656899e+05</td>\n",
       "      <td>4.670176e+02</td>\n",
       "      <td>194</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>toyota</th>\n",
       "      <td>9.018001e+04</td>\n",
       "      <td>5.131922e+05</td>\n",
       "      <td>3.382641e+03</td>\n",
       "      <td>96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volkswagen</th>\n",
       "      <td>8.363219e+04</td>\n",
       "      <td>3.856171e+05</td>\n",
       "      <td>1.319960e+02</td>\n",
       "      <td>113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>volvo</th>\n",
       "      <td>2.697714e+05</td>\n",
       "      <td>4.329832e+05</td>\n",
       "      <td>1.587133e+05</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            mean           max           min  oem\n",
       "audi                2.138273e+05  5.108954e+05  1.461933e+04   28\n",
       "bmw                 2.453293e+05  1.717239e+06  1.580588e+02   27\n",
       "chevrolet           4.489312e+04  2.416497e+05  4.295486e+02   93\n",
       "datsun              4.808099e+04  1.325812e+05  3.669611e+02   47\n",
       "fiat                4.171439e+04  9.962768e+04  3.312697e+03   20\n",
       "ford                6.219472e+04  2.092850e+05  1.654592e+03  106\n",
       "honda               6.276095e+04  5.132928e+05  7.773015e+01  390\n",
       "hyundai             6.222251e+04  1.207814e+06  1.594018e+01  732\n",
       "isuzu               6.568604e+04  6.977165e+04  6.160043e+04    2\n",
       "jaguar              2.141860e+05  2.369526e+05  1.914194e+05    2\n",
       "jeep                1.171754e+05  3.292687e+05  5.813732e+02   30\n",
       "kia                 8.948121e+04  3.148552e+05  1.030224e+03   57\n",
       "land rover          4.867652e+04  4.867652e+04  4.867652e+04    1\n",
       "mahindra            1.131091e+05  1.328702e+06  6.647849e+02  212\n",
       "mahindra renault    3.280102e+04  5.241530e+04  1.387338e+04    3\n",
       "mahindra ssangyong  6.522290e+04  1.844679e+05  3.891479e+02    4\n",
       "maruti              5.379049e+04  3.395944e+05  6.581418e+01  985\n",
       "mercedes-benz       1.951546e+05  8.761309e+05  7.065350e+03   40\n",
       "mg                  1.332216e+05  3.185510e+05  5.628516e+03   26\n",
       "mini                3.478913e+05  3.478913e+05  3.478913e+05    1\n",
       "mitsubishi          1.746206e+05  6.813291e+05  1.510951e+04    9\n",
       "nissan              7.760715e+04  2.494148e+05  1.326066e+03   54\n",
       "porsche             1.541247e+06  1.541247e+06  1.541247e+06    1\n",
       "renault             6.554328e+04  2.581085e+05  1.502180e+03  119\n",
       "skoda               1.114758e+05  6.627035e+05  2.155874e+03   81\n",
       "tata                7.780391e+04  3.656899e+05  4.670176e+02  194\n",
       "toyota              9.018001e+04  5.131922e+05  3.382641e+03   96\n",
       "volkswagen          8.363219e+04  3.856171e+05  1.319960e+02  113\n",
       "volvo               2.697714e+05  4.329832e+05  1.587133e+05    4"
      ]
     },
     "execution_count": 210,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat([mean_errors, counts], axis = 1)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}