{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "96202672",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import re\n",
    "import csv\n",
    "\n",
    "from bs4 import BeautifulSoup\n",
    "import requests\n",
    "import time, os \n",
    "import random\n",
    "from fake_useragent import UserAgent\n",
    "\n",
    "pd.set_option('display.max_rows', None)\n",
    "pd.set_option('display.max_columns', None)\n",
    "\n",
    "ua = UserAgent()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6c513a08",
   "metadata": {},
   "outputs": [],
   "source": [
    "movies = pd.read_csv('./movies_2021_2017.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "56c2e0b6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4452, 12)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "4927057d",
   "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>Unnamed: 0</th>\n",
       "      <th>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4447</th>\n",
       "      <td>4447</td>\n",
       "      <td>WHAT IF? A (Fan-Made) 'Life is Strange' Story</td>\n",
       "      <td>tt7732486</td>\n",
       "      <td>https://www.imdb.com/title/tt7732486/?ref_=adv...</td>\n",
       "      <td>[]</td>\n",
       "      <td>127</td>\n",
       "      <td>Drama, Mystery</td>\n",
       "      <td>5.1</td>\n",
       "      <td>204</td>\n",
       "      <td>[]</td>\n",
       "      <td>See full summary</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4448</th>\n",
       "      <td>4448</td>\n",
       "      <td>Created Equal</td>\n",
       "      <td>tt5873100</td>\n",
       "      <td>https://www.imdb.com/title/tt5873100/?ref_=adv...</td>\n",
       "      <td>TV-PG</td>\n",
       "      <td>91</td>\n",
       "      <td>Drama, Thriller</td>\n",
       "      <td>5.5</td>\n",
       "      <td>204</td>\n",
       "      <td>[]</td>\n",
       "      <td>Bill Duke</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4449</th>\n",
       "      <td>4449</td>\n",
       "      <td>Kate Can't Swim</td>\n",
       "      <td>tt5752904</td>\n",
       "      <td>https://www.imdb.com/title/tt5752904/?ref_=adv...</td>\n",
       "      <td>[]</td>\n",
       "      <td>90</td>\n",
       "      <td>Drama</td>\n",
       "      <td>5.5</td>\n",
       "      <td>201</td>\n",
       "      <td>[]</td>\n",
       "      <td>Josh Helman</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4450</th>\n",
       "      <td>4450</td>\n",
       "      <td>Blue Hollywood</td>\n",
       "      <td>tt5588118</td>\n",
       "      <td>https://www.imdb.com/title/tt5588118/?ref_=adv...</td>\n",
       "      <td>[]</td>\n",
       "      <td>82</td>\n",
       "      <td>Comedy, Drama, Romance</td>\n",
       "      <td>7.3</td>\n",
       "      <td>200</td>\n",
       "      <td>[]</td>\n",
       "      <td>See full summary</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4451</th>\n",
       "      <td>4451</td>\n",
       "      <td>A Closer Walk with Thee</td>\n",
       "      <td>tt4943620</td>\n",
       "      <td>https://www.imdb.com/title/tt4943620/?ref_=adv...</td>\n",
       "      <td>Unrated</td>\n",
       "      <td>89</td>\n",
       "      <td>Drama, Horror</td>\n",
       "      <td>3.8</td>\n",
       "      <td>200</td>\n",
       "      <td>[]</td>\n",
       "      <td>John C. Clark</td>\n",
       "      <td>2017</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Unnamed: 0                                          title   movie_id  \\\n",
       "4447        4447  WHAT IF? A (Fan-Made) 'Life is Strange' Story  tt7732486   \n",
       "4448        4448                                  Created Equal  tt5873100   \n",
       "4449        4449                                Kate Can't Swim  tt5752904   \n",
       "4450        4450                                 Blue Hollywood  tt5588118   \n",
       "4451        4451                        A Closer Walk with Thee  tt4943620   \n",
       "\n",
       "                                        movie_imdb_link certificate  \\\n",
       "4447  https://www.imdb.com/title/tt7732486/?ref_=adv...          []   \n",
       "4448  https://www.imdb.com/title/tt5873100/?ref_=adv...       TV-PG   \n",
       "4449  https://www.imdb.com/title/tt5752904/?ref_=adv...          []   \n",
       "4450  https://www.imdb.com/title/tt5588118/?ref_=adv...          []   \n",
       "4451  https://www.imdb.com/title/tt4943620/?ref_=adv...     Unrated   \n",
       "\n",
       "     runtime_in_mins                   genre  imdb_rating  number_of_votes  \\\n",
       "4447             127          Drama, Mystery          5.1              204   \n",
       "4448              91         Drama, Thriller          5.5              204   \n",
       "4449              90                   Drama          5.5              201   \n",
       "4450              82  Comedy, Drama, Romance          7.3              200   \n",
       "4451              89           Drama, Horror          3.8              200   \n",
       "\n",
       "     metascore      top_director  release_year  \n",
       "4447        []  See full summary          2017  \n",
       "4448        []         Bill Duke          2017  \n",
       "4449        []       Josh Helman          2017  \n",
       "4450        []  See full summary          2017  \n",
       "4451        []     John C. Clark          2017  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b74138c9",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Spider-Man: No Way Home</td>\n",
       "      <td>tt10872600</td>\n",
       "      <td>https://www.imdb.com/title/tt10872600/?ref_=ad...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>148</td>\n",
       "      <td>Action, Adventure, Fantasy</td>\n",
       "      <td>8.3</td>\n",
       "      <td>739086</td>\n",
       "      <td>71</td>\n",
       "      <td>Jon Watts</td>\n",
       "      <td>2021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Dune</td>\n",
       "      <td>tt1160419</td>\n",
       "      <td>https://www.imdb.com/title/tt1160419/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>155</td>\n",
       "      <td>Action, Adventure, Drama</td>\n",
       "      <td>8.0</td>\n",
       "      <td>621069</td>\n",
       "      <td>74</td>\n",
       "      <td>Denis Villeneuve</td>\n",
       "      <td>2021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Don't Look Up</td>\n",
       "      <td>tt11286314</td>\n",
       "      <td>https://www.imdb.com/title/tt11286314/?ref_=ad...</td>\n",
       "      <td>R</td>\n",
       "      <td>138</td>\n",
       "      <td>Comedy, Drama, Sci-Fi</td>\n",
       "      <td>7.2</td>\n",
       "      <td>529981</td>\n",
       "      <td>49</td>\n",
       "      <td>Adam McKay</td>\n",
       "      <td>2021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No Time to Die</td>\n",
       "      <td>tt2382320</td>\n",
       "      <td>https://www.imdb.com/title/tt2382320/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>163</td>\n",
       "      <td>Action, Adventure, Thriller</td>\n",
       "      <td>7.3</td>\n",
       "      <td>393930</td>\n",
       "      <td>68</td>\n",
       "      <td>Cary Joji Fukunaga</td>\n",
       "      <td>2021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Zack Snyder's Justice League</td>\n",
       "      <td>tt12361974</td>\n",
       "      <td>https://www.imdb.com/title/tt12361974/?ref_=ad...</td>\n",
       "      <td>R</td>\n",
       "      <td>242</td>\n",
       "      <td>Action, Adventure, Fantasy</td>\n",
       "      <td>8.0</td>\n",
       "      <td>393504</td>\n",
       "      <td>54</td>\n",
       "      <td>Zack Snyder</td>\n",
       "      <td>2021</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          title    movie_id  \\\n",
       "0       Spider-Man: No Way Home  tt10872600   \n",
       "1                          Dune   tt1160419   \n",
       "2                 Don't Look Up  tt11286314   \n",
       "3                No Time to Die   tt2382320   \n",
       "4  Zack Snyder's Justice League  tt12361974   \n",
       "\n",
       "                                     movie_imdb_link certificate  \\\n",
       "0  https://www.imdb.com/title/tt10872600/?ref_=ad...       PG-13   \n",
       "1  https://www.imdb.com/title/tt1160419/?ref_=adv...       PG-13   \n",
       "2  https://www.imdb.com/title/tt11286314/?ref_=ad...           R   \n",
       "3  https://www.imdb.com/title/tt2382320/?ref_=adv...       PG-13   \n",
       "4  https://www.imdb.com/title/tt12361974/?ref_=ad...           R   \n",
       "\n",
       "  runtime_in_mins                        genre  imdb_rating  number_of_votes  \\\n",
       "0             148   Action, Adventure, Fantasy          8.3           739086   \n",
       "1             155     Action, Adventure, Drama          8.0           621069   \n",
       "2             138        Comedy, Drama, Sci-Fi          7.2           529981   \n",
       "3             163  Action, Adventure, Thriller          7.3           393930   \n",
       "4             242   Action, Adventure, Fantasy          8.0           393504   \n",
       "\n",
       "  metascore        top_director  release_year  \n",
       "0        71           Jon Watts          2021  \n",
       "1        74    Denis Villeneuve          2021  \n",
       "2        49          Adam McKay          2021  \n",
       "3        68  Cary Joji Fukunaga          2021  \n",
       "4        54         Zack Snyder          2021  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies = movies.drop(['Unnamed: 0'], axis=1) \n",
    "movies.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ebee96a",
   "metadata": {},
   "source": [
    "# A df with the awards link\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "3e365318",
   "metadata": {},
   "outputs": [],
   "source": [
    "movie_awards_link = movies[['title', 'movie_id', 'movie_imdb_link']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "1972621d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_53513/3436540390.py:1: 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: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link['awards_link'] = ''\n"
     ]
    }
   ],
   "source": [
    "movie_awards_link['awards_link'] = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "d62799e7",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>awards_link</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Spider-Man: No Way Home</td>\n",
       "      <td>tt10872600</td>\n",
       "      <td>https://www.imdb.com/title/tt10872600/?ref_=ad...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Dune</td>\n",
       "      <td>tt1160419</td>\n",
       "      <td>https://www.imdb.com/title/tt1160419/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Don't Look Up</td>\n",
       "      <td>tt11286314</td>\n",
       "      <td>https://www.imdb.com/title/tt11286314/?ref_=ad...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No Time to Die</td>\n",
       "      <td>tt2382320</td>\n",
       "      <td>https://www.imdb.com/title/tt2382320/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Zack Snyder's Justice League</td>\n",
       "      <td>tt12361974</td>\n",
       "      <td>https://www.imdb.com/title/tt12361974/?ref_=ad...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          title    movie_id  \\\n",
       "0       Spider-Man: No Way Home  tt10872600   \n",
       "1                          Dune   tt1160419   \n",
       "2                 Don't Look Up  tt11286314   \n",
       "3                No Time to Die   tt2382320   \n",
       "4  Zack Snyder's Justice League  tt12361974   \n",
       "\n",
       "                                     movie_imdb_link awards_link  \n",
       "0  https://www.imdb.com/title/tt10872600/?ref_=ad...              \n",
       "1  https://www.imdb.com/title/tt1160419/?ref_=adv...              \n",
       "2  https://www.imdb.com/title/tt11286314/?ref_=ad...              \n",
       "3  https://www.imdb.com/title/tt2382320/?ref_=adv...              \n",
       "4  https://www.imdb.com/title/tt12361974/?ref_=ad...              "
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_awards_link.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "d8737afb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_53513/2932944961.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link['awards_link'][i] = awards_link_temp + '/awards/?ref_=tt_awd'\n"
     ]
    }
   ],
   "source": [
    "for i in movie_awards_link.index:\n",
    "    awards_link_temp = movie_awards_link['movie_imdb_link'][i].split('/?ref_=adv_li_tt')[0]\n",
    "    movie_awards_link['awards_link'][i] = awards_link_temp + '/awards/?ref_=tt_awd'\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "02ae6ca8",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>awards_link</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Spider-Man: No Way Home</td>\n",
       "      <td>tt10872600</td>\n",
       "      <td>https://www.imdb.com/title/tt10872600/?ref_=ad...</td>\n",
       "      <td>https://www.imdb.com/title/tt10872600/awards/?...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Dune</td>\n",
       "      <td>tt1160419</td>\n",
       "      <td>https://www.imdb.com/title/tt1160419/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt1160419/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Don't Look Up</td>\n",
       "      <td>tt11286314</td>\n",
       "      <td>https://www.imdb.com/title/tt11286314/?ref_=ad...</td>\n",
       "      <td>https://www.imdb.com/title/tt11286314/awards/?...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No Time to Die</td>\n",
       "      <td>tt2382320</td>\n",
       "      <td>https://www.imdb.com/title/tt2382320/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt2382320/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Zack Snyder's Justice League</td>\n",
       "      <td>tt12361974</td>\n",
       "      <td>https://www.imdb.com/title/tt12361974/?ref_=ad...</td>\n",
       "      <td>https://www.imdb.com/title/tt12361974/awards/?...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          title    movie_id  \\\n",
       "0       Spider-Man: No Way Home  tt10872600   \n",
       "1                          Dune   tt1160419   \n",
       "2                 Don't Look Up  tt11286314   \n",
       "3                No Time to Die   tt2382320   \n",
       "4  Zack Snyder's Justice League  tt12361974   \n",
       "\n",
       "                                     movie_imdb_link  \\\n",
       "0  https://www.imdb.com/title/tt10872600/?ref_=ad...   \n",
       "1  https://www.imdb.com/title/tt1160419/?ref_=adv...   \n",
       "2  https://www.imdb.com/title/tt11286314/?ref_=ad...   \n",
       "3  https://www.imdb.com/title/tt2382320/?ref_=adv...   \n",
       "4  https://www.imdb.com/title/tt12361974/?ref_=ad...   \n",
       "\n",
       "                                         awards_link  \n",
       "0  https://www.imdb.com/title/tt10872600/awards/?...  \n",
       "1  https://www.imdb.com/title/tt1160419/awards/?r...  \n",
       "2  https://www.imdb.com/title/tt11286314/awards/?...  \n",
       "3  https://www.imdb.com/title/tt2382320/awards/?r...  \n",
       "4  https://www.imdb.com/title/tt12361974/awards/?...  "
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_awards_link.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "82a9d5ef",
   "metadata": {},
   "source": [
    "# Scraping the awards information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b15ca0b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "def scrape_awards(df):\n",
    "    \n",
    "    movie_awards_dict = {}  #initializing an empty dictionary to fill with information\n",
    "\n",
    "    for i in df.index:\n",
    "        title = df['title'][i]\n",
    "        movie_id = df['movie_id'][i]\n",
    "        awards_link = df['awards_link'][i]\n",
    "        \n",
    "        user_agent = {'User-agent': ua.random} # generates a random user agent.\n",
    "        response = requests.get(awards_link,headers = user_agent).text\n",
    "        soup = BeautifulSoup(response, \"html.parser\")\n",
    "\n",
    "        # awards and nominations:\n",
    "        try:\n",
    "            awards_total = soup.select('#main > div:nth-child(1) > div > div.header > div > div')[0].get_text()\n",
    "        except:  \n",
    "            awards_total = 0\n",
    "        \n",
    "        time.sleep(.5+2*random.random()) # random time between requests.\n",
    "        \n",
    "        movie_awards_dict[df.index[i]] = [title]+[movie_id]+[awards_link]+[awards_total]\n",
    "\n",
    "    # define headers\n",
    "    headers = ['title','movie_id','awards_link', 'awards_total']\n",
    "    \n",
    "    # turn into a dataframe    \n",
    "    movie_awards_dataframe = pd.DataFrame(movie_awards_dict).T\n",
    "    movie_awards_dataframe.columns = headers\n",
    "    \n",
    "    #save the dataframe\n",
    "    #movie_awards_dataframe.to_csv('awards_2022.csv') # CHANGE THE YEAR ACC TO YOUR INTEREST\n",
    "    \n",
    "    return movie_awards_dataframe\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "id": "dd1f50d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "awards_df = scrape_awards(movie_awards_link)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "id": "674c75a4",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>awards_link</th>\n",
       "      <th>awards_total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Spider-Man: No Way Home</td>\n",
       "      <td>tt10872600</td>\n",
       "      <td>https://www.imdb.com/title/tt10872600/awards/?...</td>\n",
       "      <td>Showing all 34 wins and 66 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Dune</td>\n",
       "      <td>tt1160419</td>\n",
       "      <td>https://www.imdb.com/title/tt1160419/awards/?r...</td>\n",
       "      <td>Showing all 164 wins and 280 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Don't Look Up</td>\n",
       "      <td>tt11286314</td>\n",
       "      <td>https://www.imdb.com/title/tt11286314/awards/?...</td>\n",
       "      <td>Showing all 19 wins and 88 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>No Time to Die</td>\n",
       "      <td>tt2382320</td>\n",
       "      <td>https://www.imdb.com/title/tt2382320/awards/?r...</td>\n",
       "      <td>Showing all 35 wins and 69 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Zack Snyder's Justice League</td>\n",
       "      <td>tt12361974</td>\n",
       "      <td>https://www.imdb.com/title/tt12361974/awards/?...</td>\n",
       "      <td>Showing all 1 win and 6 nominations</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          title    movie_id  \\\n",
       "0       Spider-Man: No Way Home  tt10872600   \n",
       "1                          Dune   tt1160419   \n",
       "2                 Don't Look Up  tt11286314   \n",
       "3                No Time to Die   tt2382320   \n",
       "4  Zack Snyder's Justice League  tt12361974   \n",
       "\n",
       "                                         awards_link  \\\n",
       "0  https://www.imdb.com/title/tt10872600/awards/?...   \n",
       "1  https://www.imdb.com/title/tt1160419/awards/?r...   \n",
       "2  https://www.imdb.com/title/tt11286314/awards/?...   \n",
       "3  https://www.imdb.com/title/tt2382320/awards/?r...   \n",
       "4  https://www.imdb.com/title/tt12361974/awards/?...   \n",
       "\n",
       "                               awards_total  \n",
       "0    Showing all 34 wins and 66 nominations  \n",
       "1  Showing all 164 wins and 280 nominations  \n",
       "2    Showing all 19 wins and 88 nominations  \n",
       "3    Showing all 35 wins and 69 nominations  \n",
       "4       Showing all 1 win and 6 nominations  "
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "awards_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca1bf690",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "03301b5b",
   "metadata": {},
   "source": [
    "# Repeating everything for another dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "id": "19a0bb55",
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_2016_2015 = pd.read_csv('./movies_2016_2015.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "id": "deb424df",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1885</th>\n",
       "      <td>The Haunting at Woodland Hills</td>\n",
       "      <td>tt1523575</td>\n",
       "      <td>https://www.imdb.com/title/tt1523575/?ref_=adv...</td>\n",
       "      <td>[]</td>\n",
       "      <td>84</td>\n",
       "      <td>Drama, Horror, Thriller</td>\n",
       "      <td>3.4</td>\n",
       "      <td>201</td>\n",
       "      <td>[]</td>\n",
       "      <td>Jeremy Casper</td>\n",
       "      <td>2016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1886</th>\n",
       "      <td>Volumes of Blood: Horror Stories</td>\n",
       "      <td>tt6211920</td>\n",
       "      <td>https://www.imdb.com/title/tt6211920/?ref_=adv...</td>\n",
       "      <td>[]</td>\n",
       "      <td>118</td>\n",
       "      <td>Horror</td>\n",
       "      <td>3.9</td>\n",
       "      <td>200</td>\n",
       "      <td>[]</td>\n",
       "      <td>See full summary</td>\n",
       "      <td>2016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1887</th>\n",
       "      <td>Game of Aces</td>\n",
       "      <td>tt3687118</td>\n",
       "      <td>https://www.imdb.com/title/tt3687118/?ref_=adv...</td>\n",
       "      <td>R</td>\n",
       "      <td>97</td>\n",
       "      <td>Action, Adventure, War</td>\n",
       "      <td>4.2</td>\n",
       "      <td>200</td>\n",
       "      <td>[]</td>\n",
       "      <td>Damien Lay</td>\n",
       "      <td>2016</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 title   movie_id  \\\n",
       "1885    The Haunting at Woodland Hills  tt1523575   \n",
       "1886  Volumes of Blood: Horror Stories  tt6211920   \n",
       "1887                      Game of Aces  tt3687118   \n",
       "\n",
       "                                        movie_imdb_link certificate  \\\n",
       "1885  https://www.imdb.com/title/tt1523575/?ref_=adv...          []   \n",
       "1886  https://www.imdb.com/title/tt6211920/?ref_=adv...          []   \n",
       "1887  https://www.imdb.com/title/tt3687118/?ref_=adv...           R   \n",
       "\n",
       "     runtime_in_mins                    genre  imdb_rating  number_of_votes  \\\n",
       "1885              84  Drama, Horror, Thriller          3.4              201   \n",
       "1886             118                   Horror          3.9              200   \n",
       "1887              97   Action, Adventure, War          4.2              200   \n",
       "\n",
       "     metascore      top_director  release_year  \n",
       "1885        []     Jeremy Casper          2016  \n",
       "1886        []  See full summary          2016  \n",
       "1887        []        Damien Lay          2016  "
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies_2016_2015 = movies_2016_2015.drop(['Unnamed: 0'], axis=1) \n",
    "movies_2016_2015.tail(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "263d23a0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "id": "4b6dd478",
   "metadata": {},
   "outputs": [],
   "source": [
    "movie_awards_link_2 = movies_2016_2015[['title', 'movie_id', 'movie_imdb_link']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "id": "789480bb",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_53513/1736227159.py:1: 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: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link_2['awards_link'] = ''\n"
     ]
    }
   ],
   "source": [
    "movie_awards_link_2['awards_link'] = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "id": "63c936da",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_53513/3521889068.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link_2['awards_link'][i] = awards_link_temp_2 + '/awards/?ref_=tt_awd'\n"
     ]
    }
   ],
   "source": [
    "for i in movie_awards_link_2.index:\n",
    "    awards_link_temp_2 = movie_awards_link_2['movie_imdb_link'][i].split('/?ref_=adv_li_tt')[0]\n",
    "    movie_awards_link_2['awards_link'][i] = awards_link_temp_2 + '/awards/?ref_=tt_awd'\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "id": "8ed04200",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>awards_link</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Star Wars: Episode VII - The Force Awakens</td>\n",
       "      <td>tt2488496</td>\n",
       "      <td>https://www.imdb.com/title/tt2488496/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt2488496/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Avengers: Age of Ultron</td>\n",
       "      <td>tt2395427</td>\n",
       "      <td>https://www.imdb.com/title/tt2395427/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt2395427/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The Martian</td>\n",
       "      <td>tt3659388</td>\n",
       "      <td>https://www.imdb.com/title/tt3659388/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt3659388/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>The Revenant</td>\n",
       "      <td>tt1663202</td>\n",
       "      <td>https://www.imdb.com/title/tt1663202/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt1663202/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Inside Out</td>\n",
       "      <td>tt2096673</td>\n",
       "      <td>https://www.imdb.com/title/tt2096673/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt2096673/awards/?r...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                        title   movie_id  \\\n",
       "0  Star Wars: Episode VII - The Force Awakens  tt2488496   \n",
       "1                     Avengers: Age of Ultron  tt2395427   \n",
       "2                                 The Martian  tt3659388   \n",
       "3                                The Revenant  tt1663202   \n",
       "4                                  Inside Out  tt2096673   \n",
       "\n",
       "                                     movie_imdb_link  \\\n",
       "0  https://www.imdb.com/title/tt2488496/?ref_=adv...   \n",
       "1  https://www.imdb.com/title/tt2395427/?ref_=adv...   \n",
       "2  https://www.imdb.com/title/tt3659388/?ref_=adv...   \n",
       "3  https://www.imdb.com/title/tt1663202/?ref_=adv...   \n",
       "4  https://www.imdb.com/title/tt2096673/?ref_=adv...   \n",
       "\n",
       "                                         awards_link  \n",
       "0  https://www.imdb.com/title/tt2488496/awards/?r...  \n",
       "1  https://www.imdb.com/title/tt2395427/awards/?r...  \n",
       "2  https://www.imdb.com/title/tt3659388/awards/?r...  \n",
       "3  https://www.imdb.com/title/tt1663202/awards/?r...  \n",
       "4  https://www.imdb.com/title/tt2096673/awards/?r...  "
      ]
     },
     "execution_count": 223,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_awards_link_2.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "id": "9ace66da",
   "metadata": {},
   "outputs": [],
   "source": [
    "awards_2016_2015_df = scrape_awards(movie_awards_link_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "8fa7eaec",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>awards_link</th>\n",
       "      <th>awards_total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Star Wars: Episode VII - The Force Awakens</td>\n",
       "      <td>tt2488496</td>\n",
       "      <td>https://www.imdb.com/title/tt2488496/awards/?r...</td>\n",
       "      <td>Showing all 62 wins and 139 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Avengers: Age of Ultron</td>\n",
       "      <td>tt2395427</td>\n",
       "      <td>https://www.imdb.com/title/tt2395427/awards/?r...</td>\n",
       "      <td>Showing all 8 wins and 51 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>The Martian</td>\n",
       "      <td>tt3659388</td>\n",
       "      <td>https://www.imdb.com/title/tt3659388/awards/?r...</td>\n",
       "      <td>Showing all 40 wins and 200 nominations</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                        title   movie_id  \\\n",
       "0  Star Wars: Episode VII - The Force Awakens  tt2488496   \n",
       "1                     Avengers: Age of Ultron  tt2395427   \n",
       "2                                 The Martian  tt3659388   \n",
       "\n",
       "                                         awards_link  \\\n",
       "0  https://www.imdb.com/title/tt2488496/awards/?r...   \n",
       "1  https://www.imdb.com/title/tt2395427/awards/?r...   \n",
       "2  https://www.imdb.com/title/tt3659388/awards/?r...   \n",
       "\n",
       "                              awards_total  \n",
       "0  Showing all 62 wins and 139 nominations  \n",
       "1    Showing all 8 wins and 51 nominations  \n",
       "2  Showing all 40 wins and 200 nominations  "
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "awards_2016_2015_df.head(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ec17eb1",
   "metadata": {},
   "source": [
    "# Validation data (year 2022)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "b3ea17e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_2022 = pd.read_csv('./movies_2022.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "213dd511",
   "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>Unnamed: 0</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>472</th>\n",
       "      <td>Call for Peace</td>\n",
       "      <td>tt19511880</td>\n",
       "      <td>https://www.imdb.com/title/tt19511880/?ref_=ad...</td>\n",
       "      <td>[]</td>\n",
       "      <td>[]</td>\n",
       "      <td>Reality-TV</td>\n",
       "      <td>9.1</td>\n",
       "      <td>204.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>Diane Paloma Eskenazi</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>473</th>\n",
       "      <td>Do Not Disturb</td>\n",
       "      <td>tt10696116</td>\n",
       "      <td>https://www.imdb.com/title/tt10696116/?ref_=ad...</td>\n",
       "      <td>[]</td>\n",
       "      <td>92</td>\n",
       "      <td>Horror</td>\n",
       "      <td>7.9</td>\n",
       "      <td>204.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>John Ainslie</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>474</th>\n",
       "      <td>Matilda the Musical</td>\n",
       "      <td>tt3447590</td>\n",
       "      <td>https://www.imdb.com/title/tt3447590/?ref_=adv...</td>\n",
       "      <td>PG</td>\n",
       "      <td>117</td>\n",
       "      <td>Comedy, Drama, Family</td>\n",
       "      <td>6.1</td>\n",
       "      <td>201.0</td>\n",
       "      <td>67</td>\n",
       "      <td>Matthew Warchus</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>475</th>\n",
       "      <td>The Haunting of the Murder House</td>\n",
       "      <td>tt20861742</td>\n",
       "      <td>https://www.imdb.com/title/tt20861742/?ref_=ad...</td>\n",
       "      <td>[]</td>\n",
       "      <td>78</td>\n",
       "      <td>Horror</td>\n",
       "      <td>4.5</td>\n",
       "      <td>200.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>Brendan Rudnicki</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>476</th>\n",
       "      <td>Project Legion</td>\n",
       "      <td>tt12885770</td>\n",
       "      <td>https://www.imdb.com/title/tt12885770/?ref_=ad...</td>\n",
       "      <td>R</td>\n",
       "      <td>90</td>\n",
       "      <td>Horror, Sci-Fi, Thriller</td>\n",
       "      <td>2.2</td>\n",
       "      <td>200.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>Lance Kawas</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           Unnamed: 0    movie_id  \\\n",
       "472                    Call for Peace  tt19511880   \n",
       "473                    Do Not Disturb  tt10696116   \n",
       "474               Matilda the Musical   tt3447590   \n",
       "475  The Haunting of the Murder House  tt20861742   \n",
       "476                    Project Legion  tt12885770   \n",
       "\n",
       "                                       movie_imdb_link certificate  \\\n",
       "472  https://www.imdb.com/title/tt19511880/?ref_=ad...          []   \n",
       "473  https://www.imdb.com/title/tt10696116/?ref_=ad...          []   \n",
       "474  https://www.imdb.com/title/tt3447590/?ref_=adv...          PG   \n",
       "475  https://www.imdb.com/title/tt20861742/?ref_=ad...          []   \n",
       "476  https://www.imdb.com/title/tt12885770/?ref_=ad...           R   \n",
       "\n",
       "    runtime_in_mins                     genre  imdb_rating  number_of_votes  \\\n",
       "472              []                Reality-TV          9.1            204.0   \n",
       "473              92                    Horror          7.9            204.0   \n",
       "474             117     Comedy, Drama, Family          6.1            201.0   \n",
       "475              78                    Horror          4.5            200.0   \n",
       "476              90  Horror, Sci-Fi, Thriller          2.2            200.0   \n",
       "\n",
       "    metascore           top_director  release_year  \n",
       "472        []  Diane Paloma Eskenazi          2022  \n",
       "473        []           John Ainslie          2022  \n",
       "474        67        Matthew Warchus          2022  \n",
       "475        []       Brendan Rudnicki          2022  \n",
       "476        []            Lance Kawas          2022  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies_2022.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "3c8b2dce",
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_2022 = movies_2022.reset_index()\n",
    "movies_2022 = movies_2022.rename(columns = {'Unnamed: 0':'title'})\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "af2fd151",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(477, 12)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies_2022.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "dc725f35",
   "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>index</th>\n",
       "      <th>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>The Batman</td>\n",
       "      <td>tt1877830</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>176</td>\n",
       "      <td>Action, Crime, Drama</td>\n",
       "      <td>7.9</td>\n",
       "      <td>613958.0</td>\n",
       "      <td>72</td>\n",
       "      <td>Matt Reeves</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>Top Gun: Maverick</td>\n",
       "      <td>tt1745960</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>130</td>\n",
       "      <td>Action, Drama</td>\n",
       "      <td>8.4</td>\n",
       "      <td>413022.0</td>\n",
       "      <td>78</td>\n",
       "      <td>Joseph Kosinski</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>Doctor Strange in the Multiverse of Madness</td>\n",
       "      <td>tt9419884</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>126</td>\n",
       "      <td>Action, Adventure, Fantasy</td>\n",
       "      <td>6.9</td>\n",
       "      <td>398982.0</td>\n",
       "      <td>60</td>\n",
       "      <td>Sam Raimi</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>Thor: Love and Thunder</td>\n",
       "      <td>tt10648342</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/?ref_=ad...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>118</td>\n",
       "      <td>Action, Adventure, Comedy</td>\n",
       "      <td>6.4</td>\n",
       "      <td>296865.0</td>\n",
       "      <td>57</td>\n",
       "      <td>Taika Waititi</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>Everything Everywhere All at Once</td>\n",
       "      <td>tt6710474</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/?ref_=adv...</td>\n",
       "      <td>R</td>\n",
       "      <td>139</td>\n",
       "      <td>Action, Adventure, Comedy</td>\n",
       "      <td>8.1</td>\n",
       "      <td>228555.0</td>\n",
       "      <td>81</td>\n",
       "      <td>Dan Kwan</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index                                        title    movie_id  \\\n",
       "0      0                                   The Batman   tt1877830   \n",
       "1      1                            Top Gun: Maverick   tt1745960   \n",
       "2      2  Doctor Strange in the Multiverse of Madness   tt9419884   \n",
       "3      3                       Thor: Love and Thunder  tt10648342   \n",
       "4      4            Everything Everywhere All at Once   tt6710474   \n",
       "\n",
       "                                     movie_imdb_link certificate  \\\n",
       "0  https://www.imdb.com/title/tt1877830/?ref_=adv...       PG-13   \n",
       "1  https://www.imdb.com/title/tt1745960/?ref_=adv...       PG-13   \n",
       "2  https://www.imdb.com/title/tt9419884/?ref_=adv...       PG-13   \n",
       "3  https://www.imdb.com/title/tt10648342/?ref_=ad...       PG-13   \n",
       "4  https://www.imdb.com/title/tt6710474/?ref_=adv...           R   \n",
       "\n",
       "  runtime_in_mins                       genre  imdb_rating  number_of_votes  \\\n",
       "0             176        Action, Crime, Drama          7.9         613958.0   \n",
       "1             130               Action, Drama          8.4         413022.0   \n",
       "2             126  Action, Adventure, Fantasy          6.9         398982.0   \n",
       "3             118   Action, Adventure, Comedy          6.4         296865.0   \n",
       "4             139   Action, Adventure, Comedy          8.1         228555.0   \n",
       "\n",
       "  metascore     top_director  release_year  \n",
       "0        72      Matt Reeves          2022  \n",
       "1        78  Joseph Kosinski          2022  \n",
       "2        60        Sam Raimi          2022  \n",
       "3        57    Taika Waititi          2022  \n",
       "4        81         Dan Kwan          2022  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies_2022.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "b467a05b",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>certificate</th>\n",
       "      <th>runtime_in_mins</th>\n",
       "      <th>genre</th>\n",
       "      <th>imdb_rating</th>\n",
       "      <th>number_of_votes</th>\n",
       "      <th>metascore</th>\n",
       "      <th>top_director</th>\n",
       "      <th>release_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The Batman</td>\n",
       "      <td>tt1877830</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>176</td>\n",
       "      <td>Action, Crime, Drama</td>\n",
       "      <td>7.9</td>\n",
       "      <td>613958.0</td>\n",
       "      <td>72</td>\n",
       "      <td>Matt Reeves</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Top Gun: Maverick</td>\n",
       "      <td>tt1745960</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>130</td>\n",
       "      <td>Action, Drama</td>\n",
       "      <td>8.4</td>\n",
       "      <td>413022.0</td>\n",
       "      <td>78</td>\n",
       "      <td>Joseph Kosinski</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Doctor Strange in the Multiverse of Madness</td>\n",
       "      <td>tt9419884</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/?ref_=adv...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>126</td>\n",
       "      <td>Action, Adventure, Fantasy</td>\n",
       "      <td>6.9</td>\n",
       "      <td>398982.0</td>\n",
       "      <td>60</td>\n",
       "      <td>Sam Raimi</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Thor: Love and Thunder</td>\n",
       "      <td>tt10648342</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/?ref_=ad...</td>\n",
       "      <td>PG-13</td>\n",
       "      <td>118</td>\n",
       "      <td>Action, Adventure, Comedy</td>\n",
       "      <td>6.4</td>\n",
       "      <td>296865.0</td>\n",
       "      <td>57</td>\n",
       "      <td>Taika Waititi</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Everything Everywhere All at Once</td>\n",
       "      <td>tt6710474</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/?ref_=adv...</td>\n",
       "      <td>R</td>\n",
       "      <td>139</td>\n",
       "      <td>Action, Adventure, Comedy</td>\n",
       "      <td>8.1</td>\n",
       "      <td>228555.0</td>\n",
       "      <td>81</td>\n",
       "      <td>Dan Kwan</td>\n",
       "      <td>2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         title    movie_id  \\\n",
       "0                                   The Batman   tt1877830   \n",
       "1                            Top Gun: Maverick   tt1745960   \n",
       "2  Doctor Strange in the Multiverse of Madness   tt9419884   \n",
       "3                       Thor: Love and Thunder  tt10648342   \n",
       "4            Everything Everywhere All at Once   tt6710474   \n",
       "\n",
       "                                     movie_imdb_link certificate  \\\n",
       "0  https://www.imdb.com/title/tt1877830/?ref_=adv...       PG-13   \n",
       "1  https://www.imdb.com/title/tt1745960/?ref_=adv...       PG-13   \n",
       "2  https://www.imdb.com/title/tt9419884/?ref_=adv...       PG-13   \n",
       "3  https://www.imdb.com/title/tt10648342/?ref_=ad...       PG-13   \n",
       "4  https://www.imdb.com/title/tt6710474/?ref_=adv...           R   \n",
       "\n",
       "  runtime_in_mins                       genre  imdb_rating  number_of_votes  \\\n",
       "0             176        Action, Crime, Drama          7.9         613958.0   \n",
       "1             130               Action, Drama          8.4         413022.0   \n",
       "2             126  Action, Adventure, Fantasy          6.9         398982.0   \n",
       "3             118   Action, Adventure, Comedy          6.4         296865.0   \n",
       "4             139   Action, Adventure, Comedy          8.1         228555.0   \n",
       "\n",
       "  metascore     top_director  release_year  \n",
       "0        72      Matt Reeves          2022  \n",
       "1        78  Joseph Kosinski          2022  \n",
       "2        60        Sam Raimi          2022  \n",
       "3        57    Taika Waititi          2022  \n",
       "4        81         Dan Kwan          2022  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movies_2022 = movies_2022.drop(['index'], axis=1) \n",
    "movies_2022.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "246401cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "movie_awards_link_2022 = movies_2022[['title', 'movie_id', 'movie_imdb_link']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "5accd1ae",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_61402/3446922147.py:1: 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: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link_2022['awards_link'] = ''\n"
     ]
    }
   ],
   "source": [
    "movie_awards_link_2022['awards_link'] = ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "6bfa3c64",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>awards_link</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The Batman</td>\n",
       "      <td>tt1877830</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Top Gun: Maverick</td>\n",
       "      <td>tt1745960</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Doctor Strange in the Multiverse of Madness</td>\n",
       "      <td>tt9419884</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Thor: Love and Thunder</td>\n",
       "      <td>tt10648342</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/?ref_=ad...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Everything Everywhere All at Once</td>\n",
       "      <td>tt6710474</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/?ref_=adv...</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         title    movie_id  \\\n",
       "0                                   The Batman   tt1877830   \n",
       "1                            Top Gun: Maverick   tt1745960   \n",
       "2  Doctor Strange in the Multiverse of Madness   tt9419884   \n",
       "3                       Thor: Love and Thunder  tt10648342   \n",
       "4            Everything Everywhere All at Once   tt6710474   \n",
       "\n",
       "                                     movie_imdb_link awards_link  \n",
       "0  https://www.imdb.com/title/tt1877830/?ref_=adv...              \n",
       "1  https://www.imdb.com/title/tt1745960/?ref_=adv...              \n",
       "2  https://www.imdb.com/title/tt9419884/?ref_=adv...              \n",
       "3  https://www.imdb.com/title/tt10648342/?ref_=ad...              \n",
       "4  https://www.imdb.com/title/tt6710474/?ref_=adv...              "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_awards_link_2022.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2795a252",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/33/p_3l01b14g96rn22vzwly2g00000gn/T/ipykernel_61402/3228027479.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  movie_awards_link_2022['awards_link'][i] = awards_link_temp_2022 + '/awards/?ref_=tt_awd'\n"
     ]
    }
   ],
   "source": [
    "for i in movie_awards_link_2022.index:\n",
    "    awards_link_temp_2022 = movie_awards_link_2022['movie_imdb_link'][i].split('/?ref_=adv_li_tt')[0]\n",
    "    movie_awards_link_2022['awards_link'][i] = awards_link_temp_2022 + '/awards/?ref_=tt_awd'\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "cecc4490",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>movie_imdb_link</th>\n",
       "      <th>awards_link</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The Batman</td>\n",
       "      <td>tt1877830</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Top Gun: Maverick</td>\n",
       "      <td>tt1745960</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Doctor Strange in the Multiverse of Madness</td>\n",
       "      <td>tt9419884</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/awards/?r...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Thor: Love and Thunder</td>\n",
       "      <td>tt10648342</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/?ref_=ad...</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/awards/?...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Everything Everywhere All at Once</td>\n",
       "      <td>tt6710474</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/?ref_=adv...</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/awards/?r...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         title    movie_id  \\\n",
       "0                                   The Batman   tt1877830   \n",
       "1                            Top Gun: Maverick   tt1745960   \n",
       "2  Doctor Strange in the Multiverse of Madness   tt9419884   \n",
       "3                       Thor: Love and Thunder  tt10648342   \n",
       "4            Everything Everywhere All at Once   tt6710474   \n",
       "\n",
       "                                     movie_imdb_link  \\\n",
       "0  https://www.imdb.com/title/tt1877830/?ref_=adv...   \n",
       "1  https://www.imdb.com/title/tt1745960/?ref_=adv...   \n",
       "2  https://www.imdb.com/title/tt9419884/?ref_=adv...   \n",
       "3  https://www.imdb.com/title/tt10648342/?ref_=ad...   \n",
       "4  https://www.imdb.com/title/tt6710474/?ref_=adv...   \n",
       "\n",
       "                                         awards_link  \n",
       "0  https://www.imdb.com/title/tt1877830/awards/?r...  \n",
       "1  https://www.imdb.com/title/tt1745960/awards/?r...  \n",
       "2  https://www.imdb.com/title/tt9419884/awards/?r...  \n",
       "3  https://www.imdb.com/title/tt10648342/awards/?...  \n",
       "4  https://www.imdb.com/title/tt6710474/awards/?r...  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "movie_awards_link_2022.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "1a9280e3",
   "metadata": {},
   "outputs": [],
   "source": [
    "awards_2022_df = scrape_awards(movie_awards_link_2022)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "3a0d4048",
   "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>title</th>\n",
       "      <th>movie_id</th>\n",
       "      <th>awards_link</th>\n",
       "      <th>awards_total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The Batman</td>\n",
       "      <td>tt1877830</td>\n",
       "      <td>https://www.imdb.com/title/tt1877830/awards/?r...</td>\n",
       "      <td>Showing all 2 wins and 27 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Top Gun: Maverick</td>\n",
       "      <td>tt1745960</td>\n",
       "      <td>https://www.imdb.com/title/tt1745960/awards/?r...</td>\n",
       "      <td>Showing all 2 wins and 10 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Doctor Strange in the Multiverse of Madness</td>\n",
       "      <td>tt9419884</td>\n",
       "      <td>https://www.imdb.com/title/tt9419884/awards/?r...</td>\n",
       "      <td>Showing all 0 wins and 8 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Thor: Love and Thunder</td>\n",
       "      <td>tt10648342</td>\n",
       "      <td>https://www.imdb.com/title/tt10648342/awards/?...</td>\n",
       "      <td>Showing all 0 wins and 6 nominations</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Everything Everywhere All at Once</td>\n",
       "      <td>tt6710474</td>\n",
       "      <td>https://www.imdb.com/title/tt6710474/awards/?r...</td>\n",
       "      <td>Showing all 10 wins and 14 nominations</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         title    movie_id  \\\n",
       "0                                   The Batman   tt1877830   \n",
       "1                            Top Gun: Maverick   tt1745960   \n",
       "2  Doctor Strange in the Multiverse of Madness   tt9419884   \n",
       "3                       Thor: Love and Thunder  tt10648342   \n",
       "4            Everything Everywhere All at Once   tt6710474   \n",
       "\n",
       "                                         awards_link  \\\n",
       "0  https://www.imdb.com/title/tt1877830/awards/?r...   \n",
       "1  https://www.imdb.com/title/tt1745960/awards/?r...   \n",
       "2  https://www.imdb.com/title/tt9419884/awards/?r...   \n",
       "3  https://www.imdb.com/title/tt10648342/awards/?...   \n",
       "4  https://www.imdb.com/title/tt6710474/awards/?r...   \n",
       "\n",
       "                             awards_total  \n",
       "0   Showing all 2 wins and 27 nominations  \n",
       "1   Showing all 2 wins and 10 nominations  \n",
       "2    Showing all 0 wins and 8 nominations  \n",
       "3    Showing all 0 wins and 6 nominations  \n",
       "4  Showing all 10 wins and 14 nominations  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "awards_2022_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "57941250",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.12"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}