nn_try.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/stas/anaconda3/lib/python3.7/site-packages/tensorflow/python/compat/v2_compat.py:96: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "non-resource variables are not supported in the long term\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "import tensorflow.compat.v1 as tf\n",
    "tf.disable_v2_behavior() \n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import KFold\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn import metrics\n",
    "from statistics import mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_profitability(revenue, budget):\n",
    "    revenue, budget = float(revenue), float(budget)\n",
    "    profit = revenue - budget\n",
    "    return profit/budget if profit > 0 else profit/revenue\n",
    "\n",
    "def calculate_revenue(profitability, budget):\n",
    "    profitability, budget = float(profitability), float(budget)\n",
    "    return budget * (profitability + 1) if profitability > 0 else budget / (1 - profitability)\n",
    "\n",
    "def get_profitability_col(df, col_revenue, col_budget):\n",
    "    return [calculate_profitability(row[col_revenue], row[col_budget]) for i, row in df.iterrows()]\n",
    "\n",
    "def get_revenue_col(df, col_profitability, col_budget):\n",
    "    return [calculate_revenue(row[col_profitability], row[col_budget]) for i, row in df.iterrows()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_init = pd.read_csv('16 topics.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mean_absolute_percentage_error(y_true, y_pred):\n",
    "    y_true, y_pred = np.array(y_true), np.array(y_pred)\n",
    "    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100\n",
    "\n",
    "def get_metrics(y_test, y_pred):\n",
    "    return  {\n",
    "        'mae': metrics.mean_absolute_error(y_test, y_pred),\n",
    "        'mse': metrics.mean_squared_error(y_test, y_pred),\n",
    "        'rmse': np.sqrt(metrics.mean_squared_error(y_test, y_pred)),\n",
    "        'mape': mean_absolute_percentage_errorr(y_test, y_pred),\n",
    "    }\n",
    "\n",
    "def print_metrics(y_test, y_pred):\n",
    "    res = get_metrics(y_test, y_pred)\n",
    "    for key in res:\n",
    "        print(f'{key}: {res[key]}')\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepare_X_y(df):\n",
    "    dff = df[[col for col in df.columns if not 'META' in col and not 'PROCESS' in col]]\n",
    "    min_prof = dff['profitability'].min()\n",
    "    dff['profitability_moved'] = dff['profitability'] - min_prof + 1\n",
    "    dff = dff.drop(['revenue','profitability'], axis=1)\n",
    "    \n",
    "    X = dff.drop('profitability_moved', axis=1)\n",
    "    X.fillna(X.mean(), inplace=True)\n",
    "    y = dff['profitability_moved']\n",
    "    \n",
    "    return X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepare_X_y_revenue(df):\n",
    "    dff = df[[col for col in df.columns if not 'META' in col and not 'PROCESS' in col]]\n",
    "    dff = dff.drop(['profitability'], axis=1)\n",
    "    \n",
    "    X = dff.drop('revenue', axis=1)\n",
    "    X.fillna(X.mean(), inplace=True)\n",
    "    y = dff['revenue']\n",
    "    \n",
    "    return X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_RF_splits(X, y, splits):\n",
    "    metrics_dict = {key: [] for key in ['mae', 'mse', 'rmse', 'mape']}\n",
    "    skf = KFold(n_splits=splits)\n",
    "    split_num=0\n",
    "    for train_index, test_index in skf.split(X, y):\n",
    "        split_num += 1\n",
    "        print(f'split: {split_num}')\n",
    "        X_train, X_test = X.iloc[train_index], X.iloc[test_index]\n",
    "        y_train, y_test = y[train_index], y[test_index]\n",
    "\n",
    "        RFR = RandomForestRegressor(n_estimators=100, min_samples_split=3, random_state=0, n_jobs=-1)\n",
    "        RFR.fit(X_train, y_train)\n",
    "        y_pred = RFR.predict(X_test)\n",
    "\n",
    "        validation = print_metrics(y_test, y_pred)\n",
    "\n",
    "        for v in validation:\n",
    "            metrics_dict[v].append(validation[v])\n",
    "        print('\\n')\n",
    "\n",
    "    for v in metrics_dict:\n",
    "        print(f'mean {v}: {mean(metrics_dict[v])}')\n",
    "    return y_pred, metrics_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# run full dataset with splits\n",
    "X, y = prepare_X_y(df_init)\n",
    "full_y_pred, full_metrics = run_RF(X, y, splits=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/stas/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "split: 1\n",
      "mae: 12.353108394734958\n",
      "mse: 1135.4465362717951\n",
      "rmse: 33.69638758489989\n",
      "mape: 1.366217992298707\n",
      "\n",
      "\n",
      "split: 2\n",
      "mae: 8.597917243645593\n",
      "mse: 339.45564616756747\n",
      "rmse: 18.424322135904145\n",
      "mape: 0.9603231949601524\n",
      "\n",
      "\n",
      "split: 3\n",
      "mae: 7.61737384464277\n",
      "mse: 448.93301681510394\n",
      "rmse: 21.188039475494282\n",
      "mape: 0.8478932255723619\n",
      "\n",
      "\n",
      "split: 4\n",
      "mae: 8.217579051916989\n",
      "mse: 647.5009548990606\n",
      "rmse: 25.446040063221243\n",
      "mape: 0.886341282014275\n",
      "\n",
      "\n",
      "split: 5\n",
      "mae: 6.889201020133799\n",
      "mse: 276.26241083198073\n",
      "rmse: 16.621143487497505\n",
      "mape: 0.7736321784616385\n",
      "\n",
      "\n",
      "split: 6\n",
      "mae: 9.224845120346592\n",
      "mse: 528.195455200131\n",
      "rmse: 22.982503240511704\n",
      "mape: 1.066119129624807\n",
      "\n",
      "\n",
      "split: 7\n",
      "mae: 11.922782653824356\n",
      "mse: 1764.3238197577384\n",
      "rmse: 42.00385482021547\n",
      "mape: 1.8070498466349358\n",
      "\n",
      "\n",
      "split: 8\n",
      "mae: 17.61052927205968\n",
      "mse: 3778.7271090350496\n",
      "rmse: 61.47135193759\n",
      "mape: 177.5612496885289\n",
      "\n",
      "\n",
      "split: 9\n",
      "mae: 11.423137619153719\n",
      "mse: 687.5564418626388\n",
      "rmse: 26.221297486254162\n",
      "mape: 1.33527331924161\n",
      "\n",
      "\n",
      "split: 10\n",
      "mae: 20.010325922457884\n",
      "mse: 4169.06897853139\n",
      "rmse: 64.56832798308618\n",
      "mape: 5.00808253040513\n",
      "\n",
      "\n",
      "split: 11\n",
      "mae: 16.172645336103535\n",
      "mse: 2293.900033053122\n",
      "rmse: 47.89467645838233\n",
      "mape: 2.503026169149842\n",
      "\n",
      "\n",
      "split: 12\n",
      "mae: 11.58826661535319\n",
      "mse: 879.8979245168831\n",
      "rmse: 29.663073416571034\n",
      "mape: 1.4312134851326335\n",
      "\n",
      "\n",
      "split: 13\n",
      "mae: 14.761673888676249\n",
      "mse: 1763.1568196058433\n",
      "rmse: 41.98996093837006\n",
      "mape: 1.8215818641805537\n",
      "\n",
      "\n",
      "split: 14\n",
      "mae: 17.649852431758344\n",
      "mse: 4232.326502439009\n",
      "rmse: 65.05633329998709\n",
      "mape: 3.3859192029138097\n",
      "\n",
      "\n",
      "split: 15\n",
      "mae: 15.945469685292588\n",
      "mse: 1303.3990645727733\n",
      "rmse: 36.10261852792361\n",
      "mape: 1.9309199809481734\n",
      "\n",
      "\n",
      "mean mae: 12.66564720667335\n",
      "mean mse: 1616.5433809040057\n",
      "mean rmse: 36.88866205706058\n",
      "mean mape: 13.512322872671168\n"
     ]
    }
   ],
   "source": [
    "# run dataset without extremes with splits\n",
    "df_without_extremes = df_init.loc[(df_init.profitability < 1000) & (df_init.profitability > -1000)].reset_index(drop=True)\n",
    "X, y = prepare_X_y(df_without_extremes)\n",
    "y_pred, without_extremes_metrics = run_RF_splits(X, y, splits=15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/stas/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "split: 1\n",
      "mae: 11.08249938567679\n",
      "mse: 893.2698357742238\n",
      "rmse: 29.887620108905022\n",
      "mape: 1.1925120938581122\n",
      "\n",
      "\n",
      "split: 2\n",
      "mae: 6.828108396328576\n",
      "mse: 213.70049773850744\n",
      "rmse: 14.618498477562852\n",
      "mape: 0.7768119090636273\n",
      "\n",
      "\n",
      "split: 3\n",
      "mae: 8.584473412543147\n",
      "mse: 942.6786609742697\n",
      "rmse: 30.703072500553912\n",
      "mape: 0.9116273533086406\n",
      "\n",
      "\n",
      "split: 4\n",
      "mae: 7.527343998258807\n",
      "mse: 405.7135460348314\n",
      "rmse: 20.14233218956612\n",
      "mape: 0.8604458913788755\n",
      "\n",
      "\n",
      "split: 5\n",
      "mae: 8.130501210608445\n",
      "mse: 323.0738768014977\n",
      "rmse: 17.974255945699053\n",
      "mape: 0.9487204470599803\n",
      "\n",
      "\n",
      "split: 6\n",
      "mae: 13.747660469191786\n",
      "mse: 2335.807907610154\n",
      "rmse: 48.33019664361147\n",
      "mape: 2.3064985480560516\n",
      "\n",
      "\n",
      "split: 7\n",
      "mae: 15.608990167777568\n",
      "mse: 3508.1134902245667\n",
      "rmse: 59.22932964524051\n",
      "mape: 303.3971161708261\n",
      "\n",
      "\n",
      "split: 8\n",
      "mae: 11.956768000131042\n",
      "mse: 855.6338134180618\n",
      "rmse: 29.25121900738603\n",
      "mape: 1.4514730438478496\n",
      "\n",
      "\n",
      "split: 9\n",
      "mae: 22.232388197188893\n",
      "mse: 4595.641605850821\n",
      "rmse: 67.79116170896337\n",
      "mape: 5.0421883799812415\n",
      "\n",
      "\n",
      "split: 10\n",
      "mae: 19.595153813141348\n",
      "mse: 5030.234997079784\n",
      "rmse: 70.92414960420592\n",
      "mape: 5.423526311421059\n",
      "\n",
      "\n",
      "split: 11\n",
      "mae: 12.807669108146788\n",
      "mse: 1187.7286167105324\n",
      "rmse: 34.463438840465884\n",
      "mape: 1.5906838496696472\n",
      "\n",
      "\n",
      "split: 12\n",
      "mae: 15.883563760306785\n",
      "mse: 1586.085776814541\n",
      "rmse: 39.82569242103069\n",
      "mape: 2.070320493117251\n",
      "\n",
      "\n",
      "split: 13\n",
      "mae: 11.577596347159686\n",
      "mse: 732.3662698513115\n",
      "rmse: 27.06226653204257\n",
      "mape: 1.3455484277918797\n",
      "\n",
      "\n",
      "split: 14\n",
      "mae: 18.086938567442907\n",
      "mse: 3497.650662265216\n",
      "rmse: 59.14093897010104\n",
      "mape: 4.58050896567196\n",
      "\n",
      "\n",
      "split: 15\n",
      "mae: 15.221292069805688\n",
      "mse: 952.453149529658\n",
      "rmse: 30.861839697750653\n",
      "mape: 1.7652306630520345\n",
      "\n",
      "\n",
      "mean mae: 13.258063126913884\n",
      "mean mse: 1804.0101804451986\n",
      "mean rmse: 38.68040081953901\n",
      "mean mape: 22.244214169873622\n"
     ]
    }
   ],
   "source": [
    "# run dataset without extremes and 2000 - 2016 years with splits\n",
    "df_2000_2016 = df_init.loc[\n",
    "    (df_init.profitability < 1000) & (df_init.profitability > -1000) &\n",
    "    (df_init.META__year > 1999) & (df_init.META__year < 2017)\n",
    "].reset_index(drop=True)\n",
    "X, y = prepare_X_y(df_2000_2016)\n",
    "without_extremes_2000_2016_metrics = run_RF(X, y, splits=15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mae: 0.0069421859812688145\n",
      "mse: 0.0004643333584251547\n",
      "rmse: 0.021548395727412163\n",
      "mape: 0.47894776029959757\n"
     ]
    }
   ],
   "source": [
    "# run dataset without extremes without splits, validation on profitability\n",
    "df_without_extremes = df_init.loc[\n",
    "    (df_init.profitability < 1000) & (df_init.profitability > -1000)\n",
    "].reset_index(drop=True)\n",
    "X, y = prepare_X_y(df_without_extremes)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "\n",
    "RFR = RandomForestRegressor(n_estimators=100, min_samples_split=3, n_jobs=8)\n",
    "RFR.fit(X_train, y_train)\n",
    "y_pred = RFR.predict(X_test)\n",
    "\n",
    "validation = print_metrics(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/stas/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  after removing the cwd from sys.path.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predicted profitability\n",
      "mae: 11.95857582142652\n",
      "mse: 1510.2667052802956\n",
      "rmse: 38.86215003419517\n",
      "mape: 1.8445687662727128\n",
      "predicted revenue from profitability\n",
      "mae: 2649217754.8656735\n",
      "mse: 2.1404302000301167e+19\n",
      "rmse: 4626478358.352189\n",
      "mape: 51235.729547628456\n"
     ]
    }
   ],
   "source": [
    "# run dataset without extremes without splits, validation on profitability, test with revenue\n",
    "df_without_extremes = df_init.loc[\n",
    "    (df_init.profitability < 1000) & (df_init.profitability > -1000)\n",
    "].reset_index(drop=True)\n",
    "X, y = prepare_X_y(df_without_extremes)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "\n",
    "RFR = RandomForestRegressor(n_estimators=100, min_samples_split=3, n_jobs=8)\n",
    "RFR.fit(X_train, y_train)\n",
    "y_pred = RFR.predict(X_test)\n",
    "\n",
    "print(\"predicted profitability\")\n",
    "print_metrics(y_test, y_pred)\n",
    "\n",
    "print(\"predicted revenue from profitability\")\n",
    "df_test = df_without_extremes[df_without_extremes.index.isin(X_test.index)].reindex(X_test.index)\n",
    "\n",
    "min_prof = df_test['profitability'].min()\n",
    "df_test['profitability_pred'] = y_pred - 1 + min_prof\n",
    "y_pred_revenue = get_revenue_col(df_test, 'profitability_pred', 'budget')\n",
    "validation = print_metrics(list(df_test['revenue']), y_pred_revenue)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predicted revenue\n",
      "mae: 45857965.225732565\n",
      "mse: 8555634838383524.0\n",
      "rmse: 92496674.7423037\n",
      "mape: 2618.897657751647\n",
      "predicted profitability from revenue\n",
      "mae: 13.210061534540921\n",
      "mse: 2743.3991550485694\n",
      "rmse: 52.37746800914081\n",
      "mape: 69.62818195767936\n"
     ]
    }
   ],
   "source": [
    "# revenue\n",
    "df_without_extremes = df_init.loc[\n",
    "    (df_init.profitability < 1000) & (df_init.profitability > -1000)\n",
    "].reset_index(drop=True)\n",
    "X, y = prepare_X_y_revenue(df_without_extremes)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
    "\n",
    "RFR = RandomForestRegressor(n_estimators=100, min_samples_split=3, n_jobs=8)\n",
    "RFR.fit(X_train, y_train)\n",
    "y_pred = RFR.predict(X_test)\n",
    "\n",
    "print(\"predicted revenue\")\n",
    "print_metrics(y_test, y_pred)\n",
    "\n",
    "print(\"predicted profitability from revenue\")\n",
    "df_test_revenue = df_without_extremes[df_without_extremes.index.isin(X_test.index)].reindex(X_test.index)\n",
    "min_prof = df_test_revenue['profitability'].min()\n",
    "df_test_revenue['revenue_pred'] = y_pred\n",
    "df_test_revenue['profitability_moved'] = df_test_revenue['profitability'] + 1 - min_prof\n",
    "\n",
    "df_test_revenue['profitability_pred'] = get_profitability_col(df_test_revenue, 'revenue_pred', 'budget')\n",
    "df_test_revenue['profitability_pred'] = df_test_revenue['profitability_pred'] + 1 - min_prof\n",
    "validation = print_metrics(list(df_test_revenue['profitability_moved']), list(df_test_revenue['profitability_pred']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['budget',\n",
       " 'META__id',\n",
       " 'PROCESS__original_language',\n",
       " 'revenue',\n",
       " 'runtime',\n",
       " 'spoken_languages',\n",
       " 'META__year',\n",
       " 'META__month',\n",
       " 'META__day',\n",
       " 'weekend',\n",
       " 'META__collection_name',\n",
       " 'META__cast_1_name',\n",
       " 'META__cast_2_name',\n",
       " 'META__cast_3_name',\n",
       " 'META__cast_4_name',\n",
       " 'META__cast_5_name',\n",
       " 'META__cast_6_name',\n",
       " 'META__cast_7_name',\n",
       " 'META__cast_8_name',\n",
       " 'META__crew__sound__music_editor',\n",
       " 'META__crew__sound__original_music_composer',\n",
       " 'META__crew__sound__sound_designer',\n",
       " 'META__crew__sound__sound_effects_editor',\n",
       " 'META__crew__sound__sound_re_recording_mixer',\n",
       " 'META__crew__sound__supervising_sound_editor',\n",
       " 'META__crew__directing__director__1',\n",
       " 'META__crew__directing__script_supervisor',\n",
       " 'META__crew__production__casting',\n",
       " 'META__crew__production__executive_producer__1',\n",
       " 'META__crew__production__producer__1',\n",
       " 'META__crew__production__producer__2',\n",
       " 'META__crew__editing__editor__1',\n",
       " 'META__crew__costume__costume_designer',\n",
       " 'META__crew__costume__costume_supervisor',\n",
       " 'META__crew__costume__makeup_artist',\n",
       " 'META__crew__crew__stunt_coordinator',\n",
       " 'META__crew__writing__screenplay__1',\n",
       " 'META__crew__art__art_direction',\n",
       " 'META__crew__art__production_design',\n",
       " 'META__crew__art__property_master',\n",
       " 'META__crew__art__set_decoration',\n",
       " 'META__crew__visualeffects__visual_effects_supervisor',\n",
       " 'META__crew__camera__director_of_photography',\n",
       " 'META__crew__camera__steadicam_operator',\n",
       " 'META__crew__camera__still_photographer',\n",
       " 'genre__war',\n",
       " 'genre__western',\n",
       " 'genre__mystery',\n",
       " 'genre__music',\n",
       " 'genre__crime',\n",
       " 'genre__romance',\n",
       " 'genre__action',\n",
       " 'genre__adventure',\n",
       " 'genre__thriller',\n",
       " 'genre__animation',\n",
       " 'genre__family',\n",
       " 'genre__drama',\n",
       " 'genre__comedy',\n",
       " 'genre__documentary',\n",
       " 'genre__history',\n",
       " 'genre__fantasy',\n",
       " 'genre__horror',\n",
       " 'META__production_company_1',\n",
       " 'META__production_company_2',\n",
       " 'META__production_company_3',\n",
       " 'country__rs',\n",
       " 'country__su',\n",
       " 'country__do',\n",
       " 'country__bo',\n",
       " 'country__ph',\n",
       " 'country__pe',\n",
       " 'country__id',\n",
       " 'country__gr',\n",
       " 'country__hn',\n",
       " 'country__nl',\n",
       " 'country__ee',\n",
       " 'country__ug',\n",
       " 'country__uy',\n",
       " 'country__ir',\n",
       " 'country__lb',\n",
       " 'country__pk',\n",
       " 'country__ua',\n",
       " 'country__eg',\n",
       " 'country__is',\n",
       " 'country__co',\n",
       " 'country__sk',\n",
       " 'country__bw',\n",
       " 'country__es',\n",
       " 'country__se',\n",
       " 'country__jp',\n",
       " 'country__np',\n",
       " 'country__tw',\n",
       " 'country__ie',\n",
       " 'country__lu',\n",
       " 'country__us',\n",
       " 'country__vn',\n",
       " 'country__ro',\n",
       " 'country__mn',\n",
       " 'country__cl',\n",
       " 'country__hu',\n",
       " 'country__br',\n",
       " 'country__ae',\n",
       " 'country__kr',\n",
       " 'country__be',\n",
       " 'country__ca',\n",
       " 'country__sg',\n",
       " 'country__bg',\n",
       " 'country__de',\n",
       " 'country__pl',\n",
       " 'country__mv',\n",
       " 'country__cn',\n",
       " 'country__in',\n",
       " 'country__th',\n",
       " 'country__fr',\n",
       " 'country__ru',\n",
       " 'country__dk',\n",
       " 'country__af',\n",
       " 'country__xg',\n",
       " 'country__it',\n",
       " 'country__ge',\n",
       " 'country__tr',\n",
       " 'country__za',\n",
       " 'country__ve',\n",
       " 'country__lk',\n",
       " 'country__hr',\n",
       " 'country__ng',\n",
       " 'country__ar',\n",
       " 'country__ba',\n",
       " 'country__il',\n",
       " 'country__mx',\n",
       " 'country__bd',\n",
       " 'country__my',\n",
       " 'country__nz',\n",
       " 'country__no',\n",
       " 'country__cz',\n",
       " 'country__hk',\n",
       " 'country__au',\n",
       " 'country__gb',\n",
       " 'country__lv',\n",
       " 'country__fi',\n",
       " 'profitability',\n",
       " 'day_sin',\n",
       " 'day_cos',\n",
       " 'month_sin',\n",
       " 'month_cos',\n",
       " 'cast_1_avg_profit',\n",
       " 'cast_1_avg_budget',\n",
       " 'cast_1_experience',\n",
       " 'cast_1_avg_profitability',\n",
       " 'cast_1_movies_before',\n",
       " 'cast_2_avg_profit',\n",
       " 'cast_2_avg_budget',\n",
       " 'cast_2_experience',\n",
       " 'cast_2_avg_profitability',\n",
       " 'cast_2_movies_before',\n",
       " 'cast_3_avg_profit',\n",
       " 'cast_3_avg_budget',\n",
       " 'cast_3_experience',\n",
       " 'cast_3_avg_profitability',\n",
       " 'cast_3_movies_before',\n",
       " 'cast_4_avg_profit',\n",
       " 'cast_4_avg_budget',\n",
       " 'cast_4_experience',\n",
       " 'cast_4_avg_profitability',\n",
       " 'cast_4_movies_before',\n",
       " 'cast_5_avg_profit',\n",
       " 'cast_5_avg_budget',\n",
       " 'cast_5_experience',\n",
       " 'cast_5_avg_profitability',\n",
       " 'cast_5_movies_before',\n",
       " 'cast_6_avg_profit',\n",
       " 'cast_6_avg_budget',\n",
       " 'cast_6_experience',\n",
       " 'cast_6_avg_profitability',\n",
       " 'cast_6_movies_before',\n",
       " 'cast_7_avg_profit',\n",
       " 'cast_7_avg_budget',\n",
       " 'cast_7_experience',\n",
       " 'cast_7_avg_profitability',\n",
       " 'cast_7_movies_before',\n",
       " 'cast_8_avg_profit',\n",
       " 'cast_8_avg_budget',\n",
       " 'cast_8_experience',\n",
       " 'cast_8_avg_profitability',\n",
       " 'cast_8_movies_before',\n",
       " 'cast_1_male',\n",
       " 'cast_1_female',\n",
       " 'cast_2_male',\n",
       " 'cast_2_female',\n",
       " 'cast_3_male',\n",
       " 'cast_3_female',\n",
       " 'cast_4_male',\n",
       " 'cast_4_female',\n",
       " 'cast_5_male',\n",
       " 'cast_5_female',\n",
       " 'cast_6_male',\n",
       " 'cast_6_female',\n",
       " 'cast_7_male',\n",
       " 'cast_7_female',\n",
       " 'cast_8_male',\n",
       " 'cast_8_female',\n",
       " 'crew__sound__music_editor_avg_profit',\n",
       " 'crew__sound__music_editor_avg_budget',\n",
       " 'crew__sound__music_editor_movies_before',\n",
       " 'crew__sound__original_music_composer_avg_profit',\n",
       " 'crew__sound__original_music_composer_avg_budget',\n",
       " 'crew__sound__original_music_composer_movies_before',\n",
       " 'crew__sound__sound_designer_avg_profit',\n",
       " 'crew__sound__sound_designer_avg_budget',\n",
       " 'crew__sound__sound_designer_movies_before',\n",
       " 'crew__sound__sound_effects_editor_avg_profit',\n",
       " 'crew__sound__sound_effects_editor_avg_budget',\n",
       " 'crew__sound__sound_effects_editor_movies_before',\n",
       " 'crew__sound__sound_re_recording_mixer_avg_profit',\n",
       " 'crew__sound__sound_re_recording_mixer_avg_budget',\n",
       " 'crew__sound__sound_re_recording_mixer_movies_before',\n",
       " 'crew__sound__supervising_sound_editor_avg_profit',\n",
       " 'crew__sound__supervising_sound_editor_avg_budget',\n",
       " 'crew__sound__supervising_sound_editor_movies_before',\n",
       " 'crew__directing__director__1_avg_profit',\n",
       " 'crew__directing__director__1_avg_budget',\n",
       " 'crew__directing__director__1_movies_before',\n",
       " 'crew__directing__script_supervisor_avg_profit',\n",
       " 'crew__directing__script_supervisor_avg_budget',\n",
       " 'crew__directing__script_supervisor_movies_before',\n",
       " 'crew__production__casting_avg_profit',\n",
       " 'crew__production__casting_avg_budget',\n",
       " 'crew__production__casting_movies_before',\n",
       " 'crew__production__executive_producer__1_avg_profit',\n",
       " 'crew__production__executive_producer__1_avg_budget',\n",
       " 'crew__production__executive_producer__1_movies_before',\n",
       " 'crew__production__producer__1_avg_profit',\n",
       " 'crew__production__producer__1_avg_budget',\n",
       " 'crew__production__producer__1_movies_before',\n",
       " 'crew__production__producer__2_avg_profit',\n",
       " 'crew__production__producer__2_avg_budget',\n",
       " 'crew__production__producer__2_movies_before',\n",
       " 'crew__editing__editor__1_avg_profit',\n",
       " 'crew__editing__editor__1_avg_budget',\n",
       " 'crew__editing__editor__1_movies_before',\n",
       " 'crew__costume__costume_designer_avg_profit',\n",
       " 'crew__costume__costume_designer_avg_budget',\n",
       " 'crew__costume__costume_designer_movies_before',\n",
       " 'crew__costume__costume_supervisor_avg_profit',\n",
       " 'crew__costume__costume_supervisor_avg_budget',\n",
       " 'crew__costume__costume_supervisor_movies_before',\n",
       " 'crew__costume__makeup_artist_avg_profit',\n",
       " 'crew__costume__makeup_artist_avg_budget',\n",
       " 'crew__costume__makeup_artist_movies_before',\n",
       " 'crew__crew__stunt_coordinator_avg_profit',\n",
       " 'crew__crew__stunt_coordinator_avg_budget',\n",
       " 'crew__crew__stunt_coordinator_movies_before',\n",
       " 'crew__writing__screenplay__1_avg_profit',\n",
       " 'crew__writing__screenplay__1_avg_budget',\n",
       " 'crew__writing__screenplay__1_movies_before',\n",
       " 'crew__art__art_direction_avg_profit',\n",
       " 'crew__art__art_direction_avg_budget',\n",
       " 'crew__art__art_direction_movies_before',\n",
       " 'crew__art__production_design_avg_profit',\n",
       " 'crew__art__production_design_avg_budget',\n",
       " 'crew__art__production_design_movies_before',\n",
       " 'crew__art__property_master_avg_profit',\n",
       " 'crew__art__property_master_avg_budget',\n",
       " 'crew__art__property_master_movies_before',\n",
       " 'crew__art__set_decoration_avg_profit',\n",
       " 'crew__art__set_decoration_avg_budget',\n",
       " 'crew__art__set_decoration_movies_before',\n",
       " 'crew__visualeffects__visual_effects_supervisor_avg_profit',\n",
       " 'crew__visualeffects__visual_effects_supervisor_avg_budget',\n",
       " 'crew__visualeffects__visual_effects_supervisor_movies_before',\n",
       " 'crew__camera__director_of_photography_avg_profit',\n",
       " 'crew__camera__director_of_photography_avg_budget',\n",
       " 'crew__camera__director_of_photography_movies_before',\n",
       " 'crew__camera__steadicam_operator_avg_profit',\n",
       " 'crew__camera__steadicam_operator_avg_budget',\n",
       " 'crew__camera__steadicam_operator_movies_before',\n",
       " 'crew__camera__still_photographer_avg_profit',\n",
       " 'crew__camera__still_photographer_avg_budget',\n",
       " 'crew__camera__still_photographer_movies_before',\n",
       " 'cast_avg_budget',\n",
       " 'cast_avg_profit',\n",
       " 'cast_avg_experience',\n",
       " 'cast_avg_movies_before',\n",
       " 'year_avg_profitability',\n",
       " 'year_avg_revenue',\n",
       " 'production_company_1_avg_profit',\n",
       " 'production_company_1_avg_profitability',\n",
       " 'production_company_1_avg_revenue',\n",
       " 'production_company_2_avg_profit',\n",
       " 'production_company_2_avg_profitability',\n",
       " 'production_company_2_avg_revenue',\n",
       " 'production_company_3_avg_profit',\n",
       " 'production_company_3_avg_profitability',\n",
       " 'production_company_3_avg_revenue',\n",
       " 'collection_name_avg_profit',\n",
       " 'collection_name_avg_revenue',\n",
       " 'competition',\n",
       " 'rating__g',\n",
       " 'rating__pg',\n",
       " 'rating__pg-13',\n",
       " 'rating__r',\n",
       " 'rating__nc-17',\n",
       " 'topic_0',\n",
       " 'topic_1',\n",
       " 'topic_2',\n",
       " 'topic_3',\n",
       " 'topic_4',\n",
       " 'topic_5',\n",
       " 'topic_6',\n",
       " 'topic_7',\n",
       " 'topic_8',\n",
       " 'topic_9',\n",
       " 'topic_10',\n",
       " 'topic_11',\n",
       " 'topic_12',\n",
       " 'topic_13',\n",
       " 'topic_14',\n",
       " 'META__original_title',\n",
       " 'META__title',\n",
       " 'META__keywords',\n",
       " 'META__plot',\n",
       " 'META__topic_dominant',\n",
       " 'META__topic_words']"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df_init.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}