Amazing-Python-Scripts
2377 строк · 89.5 Кб
1{2"cells": [3{4"cell_type": "code",5"execution_count": 1,6"id": "1857ea93",7"metadata": {},8"outputs": [],9"source": [10"import numpy as np\n",11"import pandas as pd"12]13},14{15"cell_type": "code",16"execution_count": 2,17"id": "c936f055",18"metadata": {},19"outputs": [],20"source": [21"movies = pd.read_csv('tmdb_5000_movies.csv')\n",22"credits = pd.read_csv('tmdb_5000_credits.csv') "23]24},25{26"cell_type": "code",27"execution_count": 3,28"id": "62a46c65",29"metadata": {},30"outputs": [31{32"data": {33"text/html": [34"<div>\n",35"<style scoped>\n",36" .dataframe tbody tr th:only-of-type {\n",37" vertical-align: middle;\n",38" }\n",39"\n",40" .dataframe tbody tr th {\n",41" vertical-align: top;\n",42" }\n",43"\n",44" .dataframe thead th {\n",45" text-align: right;\n",46" }\n",47"</style>\n",48"<table border=\"1\" class=\"dataframe\">\n",49" <thead>\n",50" <tr style=\"text-align: right;\">\n",51" <th></th>\n",52" <th>budget</th>\n",53" <th>genres</th>\n",54" <th>homepage</th>\n",55" <th>id</th>\n",56" <th>keywords</th>\n",57" <th>original_language</th>\n",58" <th>original_title</th>\n",59" <th>overview</th>\n",60" <th>popularity</th>\n",61" <th>production_companies</th>\n",62" <th>production_countries</th>\n",63" <th>release_date</th>\n",64" <th>revenue</th>\n",65" <th>runtime</th>\n",66" <th>spoken_languages</th>\n",67" <th>status</th>\n",68" <th>tagline</th>\n",69" <th>title</th>\n",70" <th>vote_average</th>\n",71" <th>vote_count</th>\n",72" </tr>\n",73" </thead>\n",74" <tbody>\n",75" <tr>\n",76" <th>0</th>\n",77" <td>237000000</td>\n",78" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",79" <td>http://www.avatarmovie.com/</td>\n",80" <td>19995</td>\n",81" <td>[{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":...</td>\n",82" <td>en</td>\n",83" <td>Avatar</td>\n",84" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",85" <td>150.437577</td>\n",86" <td>[{\"name\": \"Ingenious Film Partners\", \"id\": 289...</td>\n",87" <td>[{\"iso_3166_1\": \"US\", \"name\": \"United States o...</td>\n",88" <td>2009-12-10</td>\n",89" <td>2787965087</td>\n",90" <td>162.0</td>\n",91" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}, {\"iso...</td>\n",92" <td>Released</td>\n",93" <td>Enter the World of Pandora.</td>\n",94" <td>Avatar</td>\n",95" <td>7.2</td>\n",96" <td>11800</td>\n",97" </tr>\n",98" <tr>\n",99" <th>1</th>\n",100" <td>300000000</td>\n",101" <td>[{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"...</td>\n",102" <td>http://disney.go.com/disneypictures/pirates/</td>\n",103" <td>285</td>\n",104" <td>[{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na...</td>\n",105" <td>en</td>\n",106" <td>Pirates of the Caribbean: At World's End</td>\n",107" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",108" <td>139.082615</td>\n",109" <td>[{\"name\": \"Walt Disney Pictures\", \"id\": 2}, {\"...</td>\n",110" <td>[{\"iso_3166_1\": \"US\", \"name\": \"United States o...</td>\n",111" <td>2007-05-19</td>\n",112" <td>961000000</td>\n",113" <td>169.0</td>\n",114" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",115" <td>Released</td>\n",116" <td>At the end of the world, the adventure begins.</td>\n",117" <td>Pirates of the Caribbean: At World's End</td>\n",118" <td>6.9</td>\n",119" <td>4500</td>\n",120" </tr>\n",121" <tr>\n",122" <th>2</th>\n",123" <td>245000000</td>\n",124" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",125" <td>http://www.sonypictures.com/movies/spectre/</td>\n",126" <td>206647</td>\n",127" <td>[{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name...</td>\n",128" <td>en</td>\n",129" <td>Spectre</td>\n",130" <td>A cryptic message from Bond’s past sends him o...</td>\n",131" <td>107.376788</td>\n",132" <td>[{\"name\": \"Columbia Pictures\", \"id\": 5}, {\"nam...</td>\n",133" <td>[{\"iso_3166_1\": \"GB\", \"name\": \"United Kingdom\"...</td>\n",134" <td>2015-10-26</td>\n",135" <td>880674609</td>\n",136" <td>148.0</td>\n",137" <td>[{\"iso_639_1\": \"fr\", \"name\": \"Fran\\u00e7ais\"},...</td>\n",138" <td>Released</td>\n",139" <td>A Plan No One Escapes</td>\n",140" <td>Spectre</td>\n",141" <td>6.3</td>\n",142" <td>4466</td>\n",143" </tr>\n",144" <tr>\n",145" <th>3</th>\n",146" <td>250000000</td>\n",147" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam...</td>\n",148" <td>http://www.thedarkknightrises.com/</td>\n",149" <td>49026</td>\n",150" <td>[{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,...</td>\n",151" <td>en</td>\n",152" <td>The Dark Knight Rises</td>\n",153" <td>Following the death of District Attorney Harve...</td>\n",154" <td>112.312950</td>\n",155" <td>[{\"name\": \"Legendary Pictures\", \"id\": 923}, {\"...</td>\n",156" <td>[{\"iso_3166_1\": \"US\", \"name\": \"United States o...</td>\n",157" <td>2012-07-16</td>\n",158" <td>1084939099</td>\n",159" <td>165.0</td>\n",160" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",161" <td>Released</td>\n",162" <td>The Legend Ends</td>\n",163" <td>The Dark Knight Rises</td>\n",164" <td>7.6</td>\n",165" <td>9106</td>\n",166" </tr>\n",167" <tr>\n",168" <th>4</th>\n",169" <td>260000000</td>\n",170" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",171" <td>http://movies.disney.com/john-carter</td>\n",172" <td>49529</td>\n",173" <td>[{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":...</td>\n",174" <td>en</td>\n",175" <td>John Carter</td>\n",176" <td>John Carter is a war-weary, former military ca...</td>\n",177" <td>43.926995</td>\n",178" <td>[{\"name\": \"Walt Disney Pictures\", \"id\": 2}]</td>\n",179" <td>[{\"iso_3166_1\": \"US\", \"name\": \"United States o...</td>\n",180" <td>2012-03-07</td>\n",181" <td>284139100</td>\n",182" <td>132.0</td>\n",183" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",184" <td>Released</td>\n",185" <td>Lost in our world, found in another.</td>\n",186" <td>John Carter</td>\n",187" <td>6.1</td>\n",188" <td>2124</td>\n",189" </tr>\n",190" </tbody>\n",191"</table>\n",192"</div>"193],194"text/plain": [195" budget genres \\\n",196"0 237000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",197"1 300000000 [{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"... \n",198"2 245000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",199"3 250000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam... \n",200"4 260000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",201"\n",202" homepage id \\\n",203"0 http://www.avatarmovie.com/ 19995 \n",204"1 http://disney.go.com/disneypictures/pirates/ 285 \n",205"2 http://www.sonypictures.com/movies/spectre/ 206647 \n",206"3 http://www.thedarkknightrises.com/ 49026 \n",207"4 http://movies.disney.com/john-carter 49529 \n",208"\n",209" keywords original_language \\\n",210"0 [{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":... en \n",211"1 [{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na... en \n",212"2 [{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name... en \n",213"3 [{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,... en \n",214"4 [{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":... en \n",215"\n",216" original_title \\\n",217"0 Avatar \n",218"1 Pirates of the Caribbean: At World's End \n",219"2 Spectre \n",220"3 The Dark Knight Rises \n",221"4 John Carter \n",222"\n",223" overview popularity \\\n",224"0 In the 22nd century, a paraplegic Marine is di... 150.437577 \n",225"1 Captain Barbossa, long believed to be dead, ha... 139.082615 \n",226"2 A cryptic message from Bond’s past sends him o... 107.376788 \n",227"3 Following the death of District Attorney Harve... 112.312950 \n",228"4 John Carter is a war-weary, former military ca... 43.926995 \n",229"\n",230" production_companies \\\n",231"0 [{\"name\": \"Ingenious Film Partners\", \"id\": 289... \n",232"1 [{\"name\": \"Walt Disney Pictures\", \"id\": 2}, {\"... \n",233"2 [{\"name\": \"Columbia Pictures\", \"id\": 5}, {\"nam... \n",234"3 [{\"name\": \"Legendary Pictures\", \"id\": 923}, {\"... \n",235"4 [{\"name\": \"Walt Disney Pictures\", \"id\": 2}] \n",236"\n",237" production_countries release_date revenue \\\n",238"0 [{\"iso_3166_1\": \"US\", \"name\": \"United States o... 2009-12-10 2787965087 \n",239"1 [{\"iso_3166_1\": \"US\", \"name\": \"United States o... 2007-05-19 961000000 \n",240"2 [{\"iso_3166_1\": \"GB\", \"name\": \"United Kingdom\"... 2015-10-26 880674609 \n",241"3 [{\"iso_3166_1\": \"US\", \"name\": \"United States o... 2012-07-16 1084939099 \n",242"4 [{\"iso_3166_1\": \"US\", \"name\": \"United States o... 2012-03-07 284139100 \n",243"\n",244" runtime spoken_languages status \\\n",245"0 162.0 [{\"iso_639_1\": \"en\", \"name\": \"English\"}, {\"iso... Released \n",246"1 169.0 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",247"2 148.0 [{\"iso_639_1\": \"fr\", \"name\": \"Fran\\u00e7ais\"},... Released \n",248"3 165.0 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",249"4 132.0 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",250"\n",251" tagline \\\n",252"0 Enter the World of Pandora. \n",253"1 At the end of the world, the adventure begins. \n",254"2 A Plan No One Escapes \n",255"3 The Legend Ends \n",256"4 Lost in our world, found in another. \n",257"\n",258" title vote_average vote_count \n",259"0 Avatar 7.2 11800 \n",260"1 Pirates of the Caribbean: At World's End 6.9 4500 \n",261"2 Spectre 6.3 4466 \n",262"3 The Dark Knight Rises 7.6 9106 \n",263"4 John Carter 6.1 2124 "264]265},266"execution_count": 3,267"metadata": {},268"output_type": "execute_result"269}270],271"source": [272"movies.head()"273]274},275{276"cell_type": "code",277"execution_count": 4,278"id": "3fe5f790",279"metadata": {},280"outputs": [281{282"data": {283"text/plain": [284"(4803, 20)"285]286},287"execution_count": 4,288"metadata": {},289"output_type": "execute_result"290}291],292"source": [293"movies.shape"294]295},296{297"cell_type": "code",298"execution_count": 5,299"id": "3410a6c8",300"metadata": {},301"outputs": [302{303"data": {304"text/html": [305"<div>\n",306"<style scoped>\n",307" .dataframe tbody tr th:only-of-type {\n",308" vertical-align: middle;\n",309" }\n",310"\n",311" .dataframe tbody tr th {\n",312" vertical-align: top;\n",313" }\n",314"\n",315" .dataframe thead th {\n",316" text-align: right;\n",317" }\n",318"</style>\n",319"<table border=\"1\" class=\"dataframe\">\n",320" <thead>\n",321" <tr style=\"text-align: right;\">\n",322" <th></th>\n",323" <th>movie_id</th>\n",324" <th>title</th>\n",325" <th>cast</th>\n",326" <th>crew</th>\n",327" </tr>\n",328" </thead>\n",329" <tbody>\n",330" <tr>\n",331" <th>0</th>\n",332" <td>19995</td>\n",333" <td>Avatar</td>\n",334" <td>[{\"cast_id\": 242, \"character\": \"Jake Sully\", \"...</td>\n",335" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",336" </tr>\n",337" <tr>\n",338" <th>1</th>\n",339" <td>285</td>\n",340" <td>Pirates of the Caribbean: At World's End</td>\n",341" <td>[{\"cast_id\": 4, \"character\": \"Captain Jack Spa...</td>\n",342" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",343" </tr>\n",344" <tr>\n",345" <th>2</th>\n",346" <td>206647</td>\n",347" <td>Spectre</td>\n",348" <td>[{\"cast_id\": 1, \"character\": \"James Bond\", \"cr...</td>\n",349" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",350" </tr>\n",351" <tr>\n",352" <th>3</th>\n",353" <td>49026</td>\n",354" <td>The Dark Knight Rises</td>\n",355" <td>[{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba...</td>\n",356" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",357" </tr>\n",358" <tr>\n",359" <th>4</th>\n",360" <td>49529</td>\n",361" <td>John Carter</td>\n",362" <td>[{\"cast_id\": 5, \"character\": \"John Carter\", \"c...</td>\n",363" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",364" </tr>\n",365" </tbody>\n",366"</table>\n",367"</div>"368],369"text/plain": [370" movie_id title \\\n",371"0 19995 Avatar \n",372"1 285 Pirates of the Caribbean: At World's End \n",373"2 206647 Spectre \n",374"3 49026 The Dark Knight Rises \n",375"4 49529 John Carter \n",376"\n",377" cast \\\n",378"0 [{\"cast_id\": 242, \"character\": \"Jake Sully\", \"... \n",379"1 [{\"cast_id\": 4, \"character\": \"Captain Jack Spa... \n",380"2 [{\"cast_id\": 1, \"character\": \"James Bond\", \"cr... \n",381"3 [{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba... \n",382"4 [{\"cast_id\": 5, \"character\": \"John Carter\", \"c... \n",383"\n",384" crew \n",385"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",386"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",387"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",388"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",389"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... "390]391},392"execution_count": 5,393"metadata": {},394"output_type": "execute_result"395}396],397"source": [398"credits.head()"399]400},401{402"cell_type": "code",403"execution_count": 6,404"id": "e91d93e4",405"metadata": {},406"outputs": [],407"source": [408"movies = movies.merge(credits,on='title')"409]410},411{412"cell_type": "code",413"execution_count": 7,414"id": "b491943c",415"metadata": {},416"outputs": [417{418"data": {419"text/html": [420"<div>\n",421"<style scoped>\n",422" .dataframe tbody tr th:only-of-type {\n",423" vertical-align: middle;\n",424" }\n",425"\n",426" .dataframe tbody tr th {\n",427" vertical-align: top;\n",428" }\n",429"\n",430" .dataframe thead th {\n",431" text-align: right;\n",432" }\n",433"</style>\n",434"<table border=\"1\" class=\"dataframe\">\n",435" <thead>\n",436" <tr style=\"text-align: right;\">\n",437" <th></th>\n",438" <th>budget</th>\n",439" <th>genres</th>\n",440" <th>homepage</th>\n",441" <th>id</th>\n",442" <th>keywords</th>\n",443" <th>original_language</th>\n",444" <th>original_title</th>\n",445" <th>overview</th>\n",446" <th>popularity</th>\n",447" <th>production_companies</th>\n",448" <th>...</th>\n",449" <th>runtime</th>\n",450" <th>spoken_languages</th>\n",451" <th>status</th>\n",452" <th>tagline</th>\n",453" <th>title</th>\n",454" <th>vote_average</th>\n",455" <th>vote_count</th>\n",456" <th>movie_id</th>\n",457" <th>cast</th>\n",458" <th>crew</th>\n",459" </tr>\n",460" </thead>\n",461" <tbody>\n",462" <tr>\n",463" <th>0</th>\n",464" <td>237000000</td>\n",465" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",466" <td>http://www.avatarmovie.com/</td>\n",467" <td>19995</td>\n",468" <td>[{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":...</td>\n",469" <td>en</td>\n",470" <td>Avatar</td>\n",471" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",472" <td>150.437577</td>\n",473" <td>[{\"name\": \"Ingenious Film Partners\", \"id\": 289...</td>\n",474" <td>...</td>\n",475" <td>162.0</td>\n",476" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}, {\"iso...</td>\n",477" <td>Released</td>\n",478" <td>Enter the World of Pandora.</td>\n",479" <td>Avatar</td>\n",480" <td>7.2</td>\n",481" <td>11800</td>\n",482" <td>19995</td>\n",483" <td>[{\"cast_id\": 242, \"character\": \"Jake Sully\", \"...</td>\n",484" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",485" </tr>\n",486" <tr>\n",487" <th>1</th>\n",488" <td>300000000</td>\n",489" <td>[{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"...</td>\n",490" <td>http://disney.go.com/disneypictures/pirates/</td>\n",491" <td>285</td>\n",492" <td>[{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na...</td>\n",493" <td>en</td>\n",494" <td>Pirates of the Caribbean: At World's End</td>\n",495" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",496" <td>139.082615</td>\n",497" <td>[{\"name\": \"Walt Disney Pictures\", \"id\": 2}, {\"...</td>\n",498" <td>...</td>\n",499" <td>169.0</td>\n",500" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",501" <td>Released</td>\n",502" <td>At the end of the world, the adventure begins.</td>\n",503" <td>Pirates of the Caribbean: At World's End</td>\n",504" <td>6.9</td>\n",505" <td>4500</td>\n",506" <td>285</td>\n",507" <td>[{\"cast_id\": 4, \"character\": \"Captain Jack Spa...</td>\n",508" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",509" </tr>\n",510" <tr>\n",511" <th>2</th>\n",512" <td>245000000</td>\n",513" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",514" <td>http://www.sonypictures.com/movies/spectre/</td>\n",515" <td>206647</td>\n",516" <td>[{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name...</td>\n",517" <td>en</td>\n",518" <td>Spectre</td>\n",519" <td>A cryptic message from Bond’s past sends him o...</td>\n",520" <td>107.376788</td>\n",521" <td>[{\"name\": \"Columbia Pictures\", \"id\": 5}, {\"nam...</td>\n",522" <td>...</td>\n",523" <td>148.0</td>\n",524" <td>[{\"iso_639_1\": \"fr\", \"name\": \"Fran\\u00e7ais\"},...</td>\n",525" <td>Released</td>\n",526" <td>A Plan No One Escapes</td>\n",527" <td>Spectre</td>\n",528" <td>6.3</td>\n",529" <td>4466</td>\n",530" <td>206647</td>\n",531" <td>[{\"cast_id\": 1, \"character\": \"James Bond\", \"cr...</td>\n",532" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",533" </tr>\n",534" <tr>\n",535" <th>3</th>\n",536" <td>250000000</td>\n",537" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam...</td>\n",538" <td>http://www.thedarkknightrises.com/</td>\n",539" <td>49026</td>\n",540" <td>[{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,...</td>\n",541" <td>en</td>\n",542" <td>The Dark Knight Rises</td>\n",543" <td>Following the death of District Attorney Harve...</td>\n",544" <td>112.312950</td>\n",545" <td>[{\"name\": \"Legendary Pictures\", \"id\": 923}, {\"...</td>\n",546" <td>...</td>\n",547" <td>165.0</td>\n",548" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",549" <td>Released</td>\n",550" <td>The Legend Ends</td>\n",551" <td>The Dark Knight Rises</td>\n",552" <td>7.6</td>\n",553" <td>9106</td>\n",554" <td>49026</td>\n",555" <td>[{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba...</td>\n",556" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",557" </tr>\n",558" <tr>\n",559" <th>4</th>\n",560" <td>260000000</td>\n",561" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",562" <td>http://movies.disney.com/john-carter</td>\n",563" <td>49529</td>\n",564" <td>[{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":...</td>\n",565" <td>en</td>\n",566" <td>John Carter</td>\n",567" <td>John Carter is a war-weary, former military ca...</td>\n",568" <td>43.926995</td>\n",569" <td>[{\"name\": \"Walt Disney Pictures\", \"id\": 2}]</td>\n",570" <td>...</td>\n",571" <td>132.0</td>\n",572" <td>[{\"iso_639_1\": \"en\", \"name\": \"English\"}]</td>\n",573" <td>Released</td>\n",574" <td>Lost in our world, found in another.</td>\n",575" <td>John Carter</td>\n",576" <td>6.1</td>\n",577" <td>2124</td>\n",578" <td>49529</td>\n",579" <td>[{\"cast_id\": 5, \"character\": \"John Carter\", \"c...</td>\n",580" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",581" </tr>\n",582" </tbody>\n",583"</table>\n",584"<p>5 rows × 23 columns</p>\n",585"</div>"586],587"text/plain": [588" budget genres \\\n",589"0 237000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",590"1 300000000 [{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"... \n",591"2 245000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",592"3 250000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam... \n",593"4 260000000 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",594"\n",595" homepage id \\\n",596"0 http://www.avatarmovie.com/ 19995 \n",597"1 http://disney.go.com/disneypictures/pirates/ 285 \n",598"2 http://www.sonypictures.com/movies/spectre/ 206647 \n",599"3 http://www.thedarkknightrises.com/ 49026 \n",600"4 http://movies.disney.com/john-carter 49529 \n",601"\n",602" keywords original_language \\\n",603"0 [{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":... en \n",604"1 [{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na... en \n",605"2 [{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name... en \n",606"3 [{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,... en \n",607"4 [{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":... en \n",608"\n",609" original_title \\\n",610"0 Avatar \n",611"1 Pirates of the Caribbean: At World's End \n",612"2 Spectre \n",613"3 The Dark Knight Rises \n",614"4 John Carter \n",615"\n",616" overview popularity \\\n",617"0 In the 22nd century, a paraplegic Marine is di... 150.437577 \n",618"1 Captain Barbossa, long believed to be dead, ha... 139.082615 \n",619"2 A cryptic message from Bond’s past sends him o... 107.376788 \n",620"3 Following the death of District Attorney Harve... 112.312950 \n",621"4 John Carter is a war-weary, former military ca... 43.926995 \n",622"\n",623" production_companies ... runtime \\\n",624"0 [{\"name\": \"Ingenious Film Partners\", \"id\": 289... ... 162.0 \n",625"1 [{\"name\": \"Walt Disney Pictures\", \"id\": 2}, {\"... ... 169.0 \n",626"2 [{\"name\": \"Columbia Pictures\", \"id\": 5}, {\"nam... ... 148.0 \n",627"3 [{\"name\": \"Legendary Pictures\", \"id\": 923}, {\"... ... 165.0 \n",628"4 [{\"name\": \"Walt Disney Pictures\", \"id\": 2}] ... 132.0 \n",629"\n",630" spoken_languages status \\\n",631"0 [{\"iso_639_1\": \"en\", \"name\": \"English\"}, {\"iso... Released \n",632"1 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",633"2 [{\"iso_639_1\": \"fr\", \"name\": \"Fran\\u00e7ais\"},... Released \n",634"3 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",635"4 [{\"iso_639_1\": \"en\", \"name\": \"English\"}] Released \n",636"\n",637" tagline \\\n",638"0 Enter the World of Pandora. \n",639"1 At the end of the world, the adventure begins. \n",640"2 A Plan No One Escapes \n",641"3 The Legend Ends \n",642"4 Lost in our world, found in another. \n",643"\n",644" title vote_average vote_count movie_id \\\n",645"0 Avatar 7.2 11800 19995 \n",646"1 Pirates of the Caribbean: At World's End 6.9 4500 285 \n",647"2 Spectre 6.3 4466 206647 \n",648"3 The Dark Knight Rises 7.6 9106 49026 \n",649"4 John Carter 6.1 2124 49529 \n",650"\n",651" cast \\\n",652"0 [{\"cast_id\": 242, \"character\": \"Jake Sully\", \"... \n",653"1 [{\"cast_id\": 4, \"character\": \"Captain Jack Spa... \n",654"2 [{\"cast_id\": 1, \"character\": \"James Bond\", \"cr... \n",655"3 [{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba... \n",656"4 [{\"cast_id\": 5, \"character\": \"John Carter\", \"c... \n",657"\n",658" crew \n",659"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",660"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",661"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",662"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",663"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... \n",664"\n",665"[5 rows x 23 columns]"666]667},668"execution_count": 7,669"metadata": {},670"output_type": "execute_result"671}672],673"source": [674"movies.head()\n",675"# budget\n",676"# homepage\n",677"# id\n",678"# original_language\n",679"# original_title\n",680"# popularity\n",681"# production_comapny\n",682"# production_countries\n",683"# release-date(not sure)"684]685},686{687"cell_type": "code",688"execution_count": 8,689"id": "d517efd4",690"metadata": {},691"outputs": [],692"source": [693"movies = movies[['movie_id','title','overview','genres','keywords','cast','crew']]"694]695},696{697"cell_type": "code",698"execution_count": 9,699"id": "6e0d16dc",700"metadata": {},701"outputs": [702{703"data": {704"text/html": [705"<div>\n",706"<style scoped>\n",707" .dataframe tbody tr th:only-of-type {\n",708" vertical-align: middle;\n",709" }\n",710"\n",711" .dataframe tbody tr th {\n",712" vertical-align: top;\n",713" }\n",714"\n",715" .dataframe thead th {\n",716" text-align: right;\n",717" }\n",718"</style>\n",719"<table border=\"1\" class=\"dataframe\">\n",720" <thead>\n",721" <tr style=\"text-align: right;\">\n",722" <th></th>\n",723" <th>movie_id</th>\n",724" <th>title</th>\n",725" <th>overview</th>\n",726" <th>genres</th>\n",727" <th>keywords</th>\n",728" <th>cast</th>\n",729" <th>crew</th>\n",730" </tr>\n",731" </thead>\n",732" <tbody>\n",733" <tr>\n",734" <th>0</th>\n",735" <td>19995</td>\n",736" <td>Avatar</td>\n",737" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",738" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",739" <td>[{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":...</td>\n",740" <td>[{\"cast_id\": 242, \"character\": \"Jake Sully\", \"...</td>\n",741" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",742" </tr>\n",743" <tr>\n",744" <th>1</th>\n",745" <td>285</td>\n",746" <td>Pirates of the Caribbean: At World's End</td>\n",747" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",748" <td>[{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"...</td>\n",749" <td>[{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na...</td>\n",750" <td>[{\"cast_id\": 4, \"character\": \"Captain Jack Spa...</td>\n",751" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",752" </tr>\n",753" <tr>\n",754" <th>2</th>\n",755" <td>206647</td>\n",756" <td>Spectre</td>\n",757" <td>A cryptic message from Bond’s past sends him o...</td>\n",758" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",759" <td>[{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name...</td>\n",760" <td>[{\"cast_id\": 1, \"character\": \"James Bond\", \"cr...</td>\n",761" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",762" </tr>\n",763" <tr>\n",764" <th>3</th>\n",765" <td>49026</td>\n",766" <td>The Dark Knight Rises</td>\n",767" <td>Following the death of District Attorney Harve...</td>\n",768" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam...</td>\n",769" <td>[{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,...</td>\n",770" <td>[{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba...</td>\n",771" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",772" </tr>\n",773" <tr>\n",774" <th>4</th>\n",775" <td>49529</td>\n",776" <td>John Carter</td>\n",777" <td>John Carter is a war-weary, former military ca...</td>\n",778" <td>[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam...</td>\n",779" <td>[{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":...</td>\n",780" <td>[{\"cast_id\": 5, \"character\": \"John Carter\", \"c...</td>\n",781" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",782" </tr>\n",783" </tbody>\n",784"</table>\n",785"</div>"786],787"text/plain": [788" movie_id title \\\n",789"0 19995 Avatar \n",790"1 285 Pirates of the Caribbean: At World's End \n",791"2 206647 Spectre \n",792"3 49026 The Dark Knight Rises \n",793"4 49529 John Carter \n",794"\n",795" overview \\\n",796"0 In the 22nd century, a paraplegic Marine is di... \n",797"1 Captain Barbossa, long believed to be dead, ha... \n",798"2 A cryptic message from Bond’s past sends him o... \n",799"3 Following the death of District Attorney Harve... \n",800"4 John Carter is a war-weary, former military ca... \n",801"\n",802" genres \\\n",803"0 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",804"1 [{\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"... \n",805"2 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",806"3 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 80, \"nam... \n",807"4 [{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"nam... \n",808"\n",809" keywords \\\n",810"0 [{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":... \n",811"1 [{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na... \n",812"2 [{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name... \n",813"3 [{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,... \n",814"4 [{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":... \n",815"\n",816" cast \\\n",817"0 [{\"cast_id\": 242, \"character\": \"Jake Sully\", \"... \n",818"1 [{\"cast_id\": 4, \"character\": \"Captain Jack Spa... \n",819"2 [{\"cast_id\": 1, \"character\": \"James Bond\", \"cr... \n",820"3 [{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba... \n",821"4 [{\"cast_id\": 5, \"character\": \"John Carter\", \"c... \n",822"\n",823" crew \n",824"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",825"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",826"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",827"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",828"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... "829]830},831"execution_count": 9,832"metadata": {},833"output_type": "execute_result"834}835],836"source": [837"movies.head()"838]839},840{841"cell_type": "code",842"execution_count": 10,843"id": "acb58b32",844"metadata": {},845"outputs": [846{847"data": {848"text/plain": [849"movie_id 0\n",850"title 0\n",851"overview 3\n",852"genres 0\n",853"keywords 0\n",854"cast 0\n",855"crew 0\n",856"dtype: int64"857]858},859"execution_count": 10,860"metadata": {},861"output_type": "execute_result"862}863],864"source": [865"movies.isnull().sum()"866]867},868{869"cell_type": "code",870"execution_count": 11,871"id": "122e570d",872"metadata": {},873"outputs": [],874"source": [875"movies.dropna(inplace=True)"876]877},878{879"cell_type": "code",880"execution_count": 12,881"id": "135344a9",882"metadata": {},883"outputs": [],884"source": [885"#movies.duplicated().sum()"886]887},888{889"cell_type": "code",890"execution_count": 13,891"id": "be5c56eb",892"metadata": {},893"outputs": [894{895"data": {896"text/plain": [897"'[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"name\": \"Fantasy\"}, {\"id\": 878, \"name\": \"Science Fiction\"}]'"898]899},900"execution_count": 13,901"metadata": {},902"output_type": "execute_result"903}904],905"source": [906"movies.iloc[0].genres"907]908},909{910"cell_type": "code",911"execution_count": 14,912"id": "d8b61c39",913"metadata": {},914"outputs": [],915"source": [916"#ast to convert string-list to list\n",917"import ast"918]919},920{921"cell_type": "code",922"execution_count": 15,923"id": "bfcfea46",924"metadata": {},925"outputs": [],926"source": [927"def convert(text):\n",928" L = []\n",929" for i in ast.literal_eval(text):\n",930" L.append(i['name']) \n",931" return L "932]933},934{935"cell_type": "code",936"execution_count": null,937"id": "6d7af2b7",938"metadata": {},939"outputs": [],940"source": []941},942{943"cell_type": "code",944"execution_count": 16,945"id": "a643728f",946"metadata": {},947"outputs": [948{949"data": {950"text/html": [951"<div>\n",952"<style scoped>\n",953" .dataframe tbody tr th:only-of-type {\n",954" vertical-align: middle;\n",955" }\n",956"\n",957" .dataframe tbody tr th {\n",958" vertical-align: top;\n",959" }\n",960"\n",961" .dataframe thead th {\n",962" text-align: right;\n",963" }\n",964"</style>\n",965"<table border=\"1\" class=\"dataframe\">\n",966" <thead>\n",967" <tr style=\"text-align: right;\">\n",968" <th></th>\n",969" <th>movie_id</th>\n",970" <th>title</th>\n",971" <th>overview</th>\n",972" <th>genres</th>\n",973" <th>keywords</th>\n",974" <th>cast</th>\n",975" <th>crew</th>\n",976" </tr>\n",977" </thead>\n",978" <tbody>\n",979" <tr>\n",980" <th>0</th>\n",981" <td>19995</td>\n",982" <td>Avatar</td>\n",983" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",984" <td>[Action, Adventure, Fantasy, Science Fiction]</td>\n",985" <td>[{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":...</td>\n",986" <td>[{\"cast_id\": 242, \"character\": \"Jake Sully\", \"...</td>\n",987" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",988" </tr>\n",989" <tr>\n",990" <th>1</th>\n",991" <td>285</td>\n",992" <td>Pirates of the Caribbean: At World's End</td>\n",993" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",994" <td>[Adventure, Fantasy, Action]</td>\n",995" <td>[{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na...</td>\n",996" <td>[{\"cast_id\": 4, \"character\": \"Captain Jack Spa...</td>\n",997" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",998" </tr>\n",999" <tr>\n",1000" <th>2</th>\n",1001" <td>206647</td>\n",1002" <td>Spectre</td>\n",1003" <td>A cryptic message from Bond’s past sends him o...</td>\n",1004" <td>[Action, Adventure, Crime]</td>\n",1005" <td>[{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name...</td>\n",1006" <td>[{\"cast_id\": 1, \"character\": \"James Bond\", \"cr...</td>\n",1007" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",1008" </tr>\n",1009" <tr>\n",1010" <th>3</th>\n",1011" <td>49026</td>\n",1012" <td>The Dark Knight Rises</td>\n",1013" <td>Following the death of District Attorney Harve...</td>\n",1014" <td>[Action, Crime, Drama, Thriller]</td>\n",1015" <td>[{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,...</td>\n",1016" <td>[{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba...</td>\n",1017" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",1018" </tr>\n",1019" <tr>\n",1020" <th>4</th>\n",1021" <td>49529</td>\n",1022" <td>John Carter</td>\n",1023" <td>John Carter is a war-weary, former military ca...</td>\n",1024" <td>[Action, Adventure, Science Fiction]</td>\n",1025" <td>[{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":...</td>\n",1026" <td>[{\"cast_id\": 5, \"character\": \"John Carter\", \"c...</td>\n",1027" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",1028" </tr>\n",1029" </tbody>\n",1030"</table>\n",1031"</div>"1032],1033"text/plain": [1034" movie_id title \\\n",1035"0 19995 Avatar \n",1036"1 285 Pirates of the Caribbean: At World's End \n",1037"2 206647 Spectre \n",1038"3 49026 The Dark Knight Rises \n",1039"4 49529 John Carter \n",1040"\n",1041" overview \\\n",1042"0 In the 22nd century, a paraplegic Marine is di... \n",1043"1 Captain Barbossa, long believed to be dead, ha... \n",1044"2 A cryptic message from Bond’s past sends him o... \n",1045"3 Following the death of District Attorney Harve... \n",1046"4 John Carter is a war-weary, former military ca... \n",1047"\n",1048" genres \\\n",1049"0 [Action, Adventure, Fantasy, Science Fiction] \n",1050"1 [Adventure, Fantasy, Action] \n",1051"2 [Action, Adventure, Crime] \n",1052"3 [Action, Crime, Drama, Thriller] \n",1053"4 [Action, Adventure, Science Fiction] \n",1054"\n",1055" keywords \\\n",1056"0 [{\"id\": 1463, \"name\": \"culture clash\"}, {\"id\":... \n",1057"1 [{\"id\": 270, \"name\": \"ocean\"}, {\"id\": 726, \"na... \n",1058"2 [{\"id\": 470, \"name\": \"spy\"}, {\"id\": 818, \"name... \n",1059"3 [{\"id\": 849, \"name\": \"dc comics\"}, {\"id\": 853,... \n",1060"4 [{\"id\": 818, \"name\": \"based on novel\"}, {\"id\":... \n",1061"\n",1062" cast \\\n",1063"0 [{\"cast_id\": 242, \"character\": \"Jake Sully\", \"... \n",1064"1 [{\"cast_id\": 4, \"character\": \"Captain Jack Spa... \n",1065"2 [{\"cast_id\": 1, \"character\": \"James Bond\", \"cr... \n",1066"3 [{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba... \n",1067"4 [{\"cast_id\": 5, \"character\": \"John Carter\", \"c... \n",1068"\n",1069" crew \n",1070"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",1071"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",1072"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",1073"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",1074"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... "1075]1076},1077"execution_count": 16,1078"metadata": {},1079"output_type": "execute_result"1080}1081],1082"source": [1083"movies['genres'] = movies['genres'].apply(convert)\n",1084"movies.head()"1085]1086},1087{1088"cell_type": "code",1089"execution_count": 17,1090"id": "d845d787",1091"metadata": {},1092"outputs": [1093{1094"data": {1095"text/html": [1096"<div>\n",1097"<style scoped>\n",1098" .dataframe tbody tr th:only-of-type {\n",1099" vertical-align: middle;\n",1100" }\n",1101"\n",1102" .dataframe tbody tr th {\n",1103" vertical-align: top;\n",1104" }\n",1105"\n",1106" .dataframe thead th {\n",1107" text-align: right;\n",1108" }\n",1109"</style>\n",1110"<table border=\"1\" class=\"dataframe\">\n",1111" <thead>\n",1112" <tr style=\"text-align: right;\">\n",1113" <th></th>\n",1114" <th>movie_id</th>\n",1115" <th>title</th>\n",1116" <th>overview</th>\n",1117" <th>genres</th>\n",1118" <th>keywords</th>\n",1119" <th>cast</th>\n",1120" <th>crew</th>\n",1121" </tr>\n",1122" </thead>\n",1123" <tbody>\n",1124" <tr>\n",1125" <th>0</th>\n",1126" <td>19995</td>\n",1127" <td>Avatar</td>\n",1128" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",1129" <td>[Action, Adventure, Fantasy, Science Fiction]</td>\n",1130" <td>[culture clash, future, space war, space colon...</td>\n",1131" <td>[{\"cast_id\": 242, \"character\": \"Jake Sully\", \"...</td>\n",1132" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",1133" </tr>\n",1134" <tr>\n",1135" <th>1</th>\n",1136" <td>285</td>\n",1137" <td>Pirates of the Caribbean: At World's End</td>\n",1138" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",1139" <td>[Adventure, Fantasy, Action]</td>\n",1140" <td>[ocean, drug abuse, exotic island, east india ...</td>\n",1141" <td>[{\"cast_id\": 4, \"character\": \"Captain Jack Spa...</td>\n",1142" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",1143" </tr>\n",1144" <tr>\n",1145" <th>2</th>\n",1146" <td>206647</td>\n",1147" <td>Spectre</td>\n",1148" <td>A cryptic message from Bond’s past sends him o...</td>\n",1149" <td>[Action, Adventure, Crime]</td>\n",1150" <td>[spy, based on novel, secret agent, sequel, mi...</td>\n",1151" <td>[{\"cast_id\": 1, \"character\": \"James Bond\", \"cr...</td>\n",1152" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",1153" </tr>\n",1154" <tr>\n",1155" <th>3</th>\n",1156" <td>49026</td>\n",1157" <td>The Dark Knight Rises</td>\n",1158" <td>Following the death of District Attorney Harve...</td>\n",1159" <td>[Action, Crime, Drama, Thriller]</td>\n",1160" <td>[dc comics, crime fighter, terrorist, secret i...</td>\n",1161" <td>[{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba...</td>\n",1162" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",1163" </tr>\n",1164" <tr>\n",1165" <th>4</th>\n",1166" <td>49529</td>\n",1167" <td>John Carter</td>\n",1168" <td>John Carter is a war-weary, former military ca...</td>\n",1169" <td>[Action, Adventure, Science Fiction]</td>\n",1170" <td>[based on novel, mars, medallion, space travel...</td>\n",1171" <td>[{\"cast_id\": 5, \"character\": \"John Carter\", \"c...</td>\n",1172" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",1173" </tr>\n",1174" </tbody>\n",1175"</table>\n",1176"</div>"1177],1178"text/plain": [1179" movie_id title \\\n",1180"0 19995 Avatar \n",1181"1 285 Pirates of the Caribbean: At World's End \n",1182"2 206647 Spectre \n",1183"3 49026 The Dark Knight Rises \n",1184"4 49529 John Carter \n",1185"\n",1186" overview \\\n",1187"0 In the 22nd century, a paraplegic Marine is di... \n",1188"1 Captain Barbossa, long believed to be dead, ha... \n",1189"2 A cryptic message from Bond’s past sends him o... \n",1190"3 Following the death of District Attorney Harve... \n",1191"4 John Carter is a war-weary, former military ca... \n",1192"\n",1193" genres \\\n",1194"0 [Action, Adventure, Fantasy, Science Fiction] \n",1195"1 [Adventure, Fantasy, Action] \n",1196"2 [Action, Adventure, Crime] \n",1197"3 [Action, Crime, Drama, Thriller] \n",1198"4 [Action, Adventure, Science Fiction] \n",1199"\n",1200" keywords \\\n",1201"0 [culture clash, future, space war, space colon... \n",1202"1 [ocean, drug abuse, exotic island, east india ... \n",1203"2 [spy, based on novel, secret agent, sequel, mi... \n",1204"3 [dc comics, crime fighter, terrorist, secret i... \n",1205"4 [based on novel, mars, medallion, space travel... \n",1206"\n",1207" cast \\\n",1208"0 [{\"cast_id\": 242, \"character\": \"Jake Sully\", \"... \n",1209"1 [{\"cast_id\": 4, \"character\": \"Captain Jack Spa... \n",1210"2 [{\"cast_id\": 1, \"character\": \"James Bond\", \"cr... \n",1211"3 [{\"cast_id\": 2, \"character\": \"Bruce Wayne / Ba... \n",1212"4 [{\"cast_id\": 5, \"character\": \"John Carter\", \"c... \n",1213"\n",1214" crew \n",1215"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",1216"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",1217"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",1218"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",1219"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... "1220]1221},1222"execution_count": 17,1223"metadata": {},1224"output_type": "execute_result"1225}1226],1227"source": [1228"movies['keywords'] = movies['keywords'].apply(convert)\n",1229"movies.head()"1230]1231},1232{1233"cell_type": "code",1234"execution_count": 18,1235"id": "511b76bd",1236"metadata": {},1237"outputs": [1238{1239"data": {1240"text/plain": [1241"[{'id': 28, 'name': 'Action'},\n",1242" {'id': 12, 'name': 'Adventure'},\n",1243" {'id': 14, 'name': 'Fantasy'},\n",1244" {'id': 878, 'name': 'Science Fiction'}]"1245]1246},1247"execution_count": 18,1248"metadata": {},1249"output_type": "execute_result"1250}1251],1252"source": [1253"import ast\n",1254"ast.literal_eval('[{\"id\": 28, \"name\": \"Action\"}, {\"id\": 12, \"name\": \"Adventure\"}, {\"id\": 14, \"name\": \"Fantasy\"}, {\"id\": 878, \"name\":\"Science Fiction\"}]')"1255]1256},1257{1258"cell_type": "code",1259"execution_count": 19,1260"id": "7024e6a0",1261"metadata": {},1262"outputs": [],1263"source": [1264"def convert3(text):\n",1265" L = []\n",1266" counter = 0\n",1267" for i in ast.literal_eval(text):\n",1268" if counter < 3:\n",1269" L.append(i['name'])\n",1270" counter+=1\n",1271" return L "1272]1273},1274{1275"cell_type": "code",1276"execution_count": 20,1277"id": "dec536a8",1278"metadata": {},1279"outputs": [1280{1281"data": {1282"text/html": [1283"<div>\n",1284"<style scoped>\n",1285" .dataframe tbody tr th:only-of-type {\n",1286" vertical-align: middle;\n",1287" }\n",1288"\n",1289" .dataframe tbody tr th {\n",1290" vertical-align: top;\n",1291" }\n",1292"\n",1293" .dataframe thead th {\n",1294" text-align: right;\n",1295" }\n",1296"</style>\n",1297"<table border=\"1\" class=\"dataframe\">\n",1298" <thead>\n",1299" <tr style=\"text-align: right;\">\n",1300" <th></th>\n",1301" <th>movie_id</th>\n",1302" <th>title</th>\n",1303" <th>overview</th>\n",1304" <th>genres</th>\n",1305" <th>keywords</th>\n",1306" <th>cast</th>\n",1307" <th>crew</th>\n",1308" </tr>\n",1309" </thead>\n",1310" <tbody>\n",1311" <tr>\n",1312" <th>0</th>\n",1313" <td>19995</td>\n",1314" <td>Avatar</td>\n",1315" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",1316" <td>[Action, Adventure, Fantasy, Science Fiction]</td>\n",1317" <td>[culture clash, future, space war, space colon...</td>\n",1318" <td>[Sam Worthington, Zoe Saldana, Sigourney Weave...</td>\n",1319" <td>[{\"credit_id\": \"52fe48009251416c750aca23\", \"de...</td>\n",1320" </tr>\n",1321" <tr>\n",1322" <th>1</th>\n",1323" <td>285</td>\n",1324" <td>Pirates of the Caribbean: At World's End</td>\n",1325" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",1326" <td>[Adventure, Fantasy, Action]</td>\n",1327" <td>[ocean, drug abuse, exotic island, east india ...</td>\n",1328" <td>[Johnny Depp, Orlando Bloom, Keira Knightley, ...</td>\n",1329" <td>[{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de...</td>\n",1330" </tr>\n",1331" <tr>\n",1332" <th>2</th>\n",1333" <td>206647</td>\n",1334" <td>Spectre</td>\n",1335" <td>A cryptic message from Bond’s past sends him o...</td>\n",1336" <td>[Action, Adventure, Crime]</td>\n",1337" <td>[spy, based on novel, secret agent, sequel, mi...</td>\n",1338" <td>[Daniel Craig, Christoph Waltz, Léa Seydoux, R...</td>\n",1339" <td>[{\"credit_id\": \"54805967c3a36829b5002c41\", \"de...</td>\n",1340" </tr>\n",1341" <tr>\n",1342" <th>3</th>\n",1343" <td>49026</td>\n",1344" <td>The Dark Knight Rises</td>\n",1345" <td>Following the death of District Attorney Harve...</td>\n",1346" <td>[Action, Crime, Drama, Thriller]</td>\n",1347" <td>[dc comics, crime fighter, terrorist, secret i...</td>\n",1348" <td>[Christian Bale, Michael Caine, Gary Oldman, A...</td>\n",1349" <td>[{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de...</td>\n",1350" </tr>\n",1351" <tr>\n",1352" <th>4</th>\n",1353" <td>49529</td>\n",1354" <td>John Carter</td>\n",1355" <td>John Carter is a war-weary, former military ca...</td>\n",1356" <td>[Action, Adventure, Science Fiction]</td>\n",1357" <td>[based on novel, mars, medallion, space travel...</td>\n",1358" <td>[Taylor Kitsch, Lynn Collins, Samantha Morton,...</td>\n",1359" <td>[{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de...</td>\n",1360" </tr>\n",1361" </tbody>\n",1362"</table>\n",1363"</div>"1364],1365"text/plain": [1366" movie_id title \\\n",1367"0 19995 Avatar \n",1368"1 285 Pirates of the Caribbean: At World's End \n",1369"2 206647 Spectre \n",1370"3 49026 The Dark Knight Rises \n",1371"4 49529 John Carter \n",1372"\n",1373" overview \\\n",1374"0 In the 22nd century, a paraplegic Marine is di... \n",1375"1 Captain Barbossa, long believed to be dead, ha... \n",1376"2 A cryptic message from Bond’s past sends him o... \n",1377"3 Following the death of District Attorney Harve... \n",1378"4 John Carter is a war-weary, former military ca... \n",1379"\n",1380" genres \\\n",1381"0 [Action, Adventure, Fantasy, Science Fiction] \n",1382"1 [Adventure, Fantasy, Action] \n",1383"2 [Action, Adventure, Crime] \n",1384"3 [Action, Crime, Drama, Thriller] \n",1385"4 [Action, Adventure, Science Fiction] \n",1386"\n",1387" keywords \\\n",1388"0 [culture clash, future, space war, space colon... \n",1389"1 [ocean, drug abuse, exotic island, east india ... \n",1390"2 [spy, based on novel, secret agent, sequel, mi... \n",1391"3 [dc comics, crime fighter, terrorist, secret i... \n",1392"4 [based on novel, mars, medallion, space travel... \n",1393"\n",1394" cast \\\n",1395"0 [Sam Worthington, Zoe Saldana, Sigourney Weave... \n",1396"1 [Johnny Depp, Orlando Bloom, Keira Knightley, ... \n",1397"2 [Daniel Craig, Christoph Waltz, Léa Seydoux, R... \n",1398"3 [Christian Bale, Michael Caine, Gary Oldman, A... \n",1399"4 [Taylor Kitsch, Lynn Collins, Samantha Morton,... \n",1400"\n",1401" crew \n",1402"0 [{\"credit_id\": \"52fe48009251416c750aca23\", \"de... \n",1403"1 [{\"credit_id\": \"52fe4232c3a36847f800b579\", \"de... \n",1404"2 [{\"credit_id\": \"54805967c3a36829b5002c41\", \"de... \n",1405"3 [{\"credit_id\": \"52fe4781c3a36847f81398c3\", \"de... \n",1406"4 [{\"credit_id\": \"52fe479ac3a36847f813eaa3\", \"de... "1407]1408},1409"execution_count": 20,1410"metadata": {},1411"output_type": "execute_result"1412}1413],1414"source": [1415"movies['cast'] = movies['cast'].apply(convert)\n",1416"movies.head()"1417]1418},1419{1420"cell_type": "code",1421"execution_count": 21,1422"id": "ee640520",1423"metadata": {},1424"outputs": [],1425"source": [1426"movies['cast'] = movies['cast'].apply(lambda x:x[0:3])"1427]1428},1429{1430"cell_type": "code",1431"execution_count": 22,1432"id": "7597779d",1433"metadata": {},1434"outputs": [],1435"source": [1436"def fetch_director(text):\n",1437" L = []\n",1438" for i in ast.literal_eval(text):\n",1439" if i['job'] == 'Director': \n",1440" L.append(i['name'])\n",1441" return L "1442]1443},1444{1445"cell_type": "code",1446"execution_count": 23,1447"id": "913c4480",1448"metadata": {},1449"outputs": [],1450"source": [1451"movies['crew'] = movies['crew'].apply(fetch_director)"1452]1453},1454{1455"cell_type": "code",1456"execution_count": 24,1457"id": "d87dcda7",1458"metadata": {},1459"outputs": [1460{1461"data": {1462"text/html": [1463"<div>\n",1464"<style scoped>\n",1465" .dataframe tbody tr th:only-of-type {\n",1466" vertical-align: middle;\n",1467" }\n",1468"\n",1469" .dataframe tbody tr th {\n",1470" vertical-align: top;\n",1471" }\n",1472"\n",1473" .dataframe thead th {\n",1474" text-align: right;\n",1475" }\n",1476"</style>\n",1477"<table border=\"1\" class=\"dataframe\">\n",1478" <thead>\n",1479" <tr style=\"text-align: right;\">\n",1480" <th></th>\n",1481" <th>movie_id</th>\n",1482" <th>title</th>\n",1483" <th>overview</th>\n",1484" <th>genres</th>\n",1485" <th>keywords</th>\n",1486" <th>cast</th>\n",1487" <th>crew</th>\n",1488" </tr>\n",1489" </thead>\n",1490" <tbody>\n",1491" <tr>\n",1492" <th>4702</th>\n",1493" <td>79120</td>\n",1494" <td>Weekend</td>\n",1495" <td>After a drunken house party with his straight ...</td>\n",1496" <td>[Drama, Romance]</td>\n",1497" <td>[gay, great britain, one-night stand, independ...</td>\n",1498" <td>[Tom Cullen, Chris New, Jonathan Race]</td>\n",1499" <td>[Andrew Haigh]</td>\n",1500" </tr>\n",1501" <tr>\n",1502" <th>1341</th>\n",1503" <td>8470</td>\n",1504" <td>John Q</td>\n",1505" <td>John Quincy Archibald is a father and husband ...</td>\n",1506" <td>[Drama, Thriller, Crime]</td>\n",1507" <td>[father son relationship, chicago, heart attac...</td>\n",1508" <td>[Denzel Washington, Robert Duvall, Anne Heche]</td>\n",1509" <td>[Nick Cassavetes]</td>\n",1510" </tr>\n",1511" <tr>\n",1512" <th>3758</th>\n",1513" <td>8357</td>\n",1514" <td>What the #$*! Do We (K)now!?</td>\n",1515" <td>Amanda (Marlee Maitlin) is a divorced woman wh...</td>\n",1516" <td>[Documentary]</td>\n",1517" <td>[alternate dimension, new age, parallel world,...</td>\n",1518" <td>[Marlee Matlin, Elaine Hendrix, Robert Blanche]</td>\n",1519" <td>[William Arntz, Betsy Chasse]</td>\n",1520" </tr>\n",1521" <tr>\n",1522" <th>2423</th>\n",1523" <td>9792</td>\n",1524" <td>The Hills Have Eyes</td>\n",1525" <td>Based on Wes Craven's 1977 suspenseful cult cl...</td>\n",1526" <td>[Horror, Thriller]</td>\n",1527" <td>[ambush, new mexico, van, family holiday, axe ...</td>\n",1528" <td>[Aaron Stanford, Kathleen Quinlan, Vinessa Shaw]</td>\n",1529" <td>[Alexandre Aja]</td>\n",1530" </tr>\n",1531" <tr>\n",1532" <th>297</th>\n",1533" <td>1372</td>\n",1534" <td>Blood Diamond</td>\n",1535" <td>An ex-mercenary turned smuggler. A Mende fishe...</td>\n",1536" <td>[Drama, Thriller, Action]</td>\n",1537" <td>[rebel, journalist, journalism, loss of family...</td>\n",1538" <td>[Leonardo DiCaprio, Djimon Hounsou, Jennifer C...</td>\n",1539" <td>[Edward Zwick]</td>\n",1540" </tr>\n",1541" </tbody>\n",1542"</table>\n",1543"</div>"1544],1545"text/plain": [1546" movie_id title \\\n",1547"4702 79120 Weekend \n",1548"1341 8470 John Q \n",1549"3758 8357 What the #$*! Do We (K)now!? \n",1550"2423 9792 The Hills Have Eyes \n",1551"297 1372 Blood Diamond \n",1552"\n",1553" overview \\\n",1554"4702 After a drunken house party with his straight ... \n",1555"1341 John Quincy Archibald is a father and husband ... \n",1556"3758 Amanda (Marlee Maitlin) is a divorced woman wh... \n",1557"2423 Based on Wes Craven's 1977 suspenseful cult cl... \n",1558"297 An ex-mercenary turned smuggler. A Mende fishe... \n",1559"\n",1560" genres \\\n",1561"4702 [Drama, Romance] \n",1562"1341 [Drama, Thriller, Crime] \n",1563"3758 [Documentary] \n",1564"2423 [Horror, Thriller] \n",1565"297 [Drama, Thriller, Action] \n",1566"\n",1567" keywords \\\n",1568"4702 [gay, great britain, one-night stand, independ... \n",1569"1341 [father son relationship, chicago, heart attac... \n",1570"3758 [alternate dimension, new age, parallel world,... \n",1571"2423 [ambush, new mexico, van, family holiday, axe ... \n",1572"297 [rebel, journalist, journalism, loss of family... \n",1573"\n",1574" cast \\\n",1575"4702 [Tom Cullen, Chris New, Jonathan Race] \n",1576"1341 [Denzel Washington, Robert Duvall, Anne Heche] \n",1577"3758 [Marlee Matlin, Elaine Hendrix, Robert Blanche] \n",1578"2423 [Aaron Stanford, Kathleen Quinlan, Vinessa Shaw] \n",1579"297 [Leonardo DiCaprio, Djimon Hounsou, Jennifer C... \n",1580"\n",1581" crew \n",1582"4702 [Andrew Haigh] \n",1583"1341 [Nick Cassavetes] \n",1584"3758 [William Arntz, Betsy Chasse] \n",1585"2423 [Alexandre Aja] \n",1586"297 [Edward Zwick] "1587]1588},1589"execution_count": 24,1590"metadata": {},1591"output_type": "execute_result"1592}1593],1594"source": [1595"#movies['overview'] = movies['overview'].apply(lambda x:x.split())\n",1596"movies.sample(5)"1597]1598},1599{1600"cell_type": "code",1601"execution_count": 25,1602"id": "77acdc21",1603"metadata": {},1604"outputs": [],1605"source": [1606"def collapse(L):\n",1607" L1 = []\n",1608" for i in L:\n",1609" L1.append(i.replace(\" \",\"\"))\n",1610" return L1"1611]1612},1613{1614"cell_type": "code",1615"execution_count": 26,1616"id": "b3439e7c",1617"metadata": {},1618"outputs": [],1619"source": [1620"movies['cast'] = movies['cast'].apply(collapse)\n",1621"movies['crew'] = movies['crew'].apply(collapse)\n",1622"movies['genres'] = movies['genres'].apply(collapse)\n",1623"movies['keywords'] = movies['keywords'].apply(collapse)"1624]1625},1626{1627"cell_type": "code",1628"execution_count": 27,1629"id": "5684929f",1630"metadata": {},1631"outputs": [1632{1633"data": {1634"text/html": [1635"<div>\n",1636"<style scoped>\n",1637" .dataframe tbody tr th:only-of-type {\n",1638" vertical-align: middle;\n",1639" }\n",1640"\n",1641" .dataframe tbody tr th {\n",1642" vertical-align: top;\n",1643" }\n",1644"\n",1645" .dataframe thead th {\n",1646" text-align: right;\n",1647" }\n",1648"</style>\n",1649"<table border=\"1\" class=\"dataframe\">\n",1650" <thead>\n",1651" <tr style=\"text-align: right;\">\n",1652" <th></th>\n",1653" <th>movie_id</th>\n",1654" <th>title</th>\n",1655" <th>overview</th>\n",1656" <th>genres</th>\n",1657" <th>keywords</th>\n",1658" <th>cast</th>\n",1659" <th>crew</th>\n",1660" </tr>\n",1661" </thead>\n",1662" <tbody>\n",1663" <tr>\n",1664" <th>0</th>\n",1665" <td>19995</td>\n",1666" <td>Avatar</td>\n",1667" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",1668" <td>[Action, Adventure, Fantasy, ScienceFiction]</td>\n",1669" <td>[cultureclash, future, spacewar, spacecolony, ...</td>\n",1670" <td>[SamWorthington, ZoeSaldana, SigourneyWeaver]</td>\n",1671" <td>[JamesCameron]</td>\n",1672" </tr>\n",1673" <tr>\n",1674" <th>1</th>\n",1675" <td>285</td>\n",1676" <td>Pirates of the Caribbean: At World's End</td>\n",1677" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",1678" <td>[Adventure, Fantasy, Action]</td>\n",1679" <td>[ocean, drugabuse, exoticisland, eastindiatrad...</td>\n",1680" <td>[JohnnyDepp, OrlandoBloom, KeiraKnightley]</td>\n",1681" <td>[GoreVerbinski]</td>\n",1682" </tr>\n",1683" <tr>\n",1684" <th>2</th>\n",1685" <td>206647</td>\n",1686" <td>Spectre</td>\n",1687" <td>A cryptic message from Bond’s past sends him o...</td>\n",1688" <td>[Action, Adventure, Crime]</td>\n",1689" <td>[spy, basedonnovel, secretagent, sequel, mi6, ...</td>\n",1690" <td>[DanielCraig, ChristophWaltz, LéaSeydoux]</td>\n",1691" <td>[SamMendes]</td>\n",1692" </tr>\n",1693" <tr>\n",1694" <th>3</th>\n",1695" <td>49026</td>\n",1696" <td>The Dark Knight Rises</td>\n",1697" <td>Following the death of District Attorney Harve...</td>\n",1698" <td>[Action, Crime, Drama, Thriller]</td>\n",1699" <td>[dccomics, crimefighter, terrorist, secretiden...</td>\n",1700" <td>[ChristianBale, MichaelCaine, GaryOldman]</td>\n",1701" <td>[ChristopherNolan]</td>\n",1702" </tr>\n",1703" <tr>\n",1704" <th>4</th>\n",1705" <td>49529</td>\n",1706" <td>John Carter</td>\n",1707" <td>John Carter is a war-weary, former military ca...</td>\n",1708" <td>[Action, Adventure, ScienceFiction]</td>\n",1709" <td>[basedonnovel, mars, medallion, spacetravel, p...</td>\n",1710" <td>[TaylorKitsch, LynnCollins, SamanthaMorton]</td>\n",1711" <td>[AndrewStanton]</td>\n",1712" </tr>\n",1713" </tbody>\n",1714"</table>\n",1715"</div>"1716],1717"text/plain": [1718" movie_id title \\\n",1719"0 19995 Avatar \n",1720"1 285 Pirates of the Caribbean: At World's End \n",1721"2 206647 Spectre \n",1722"3 49026 The Dark Knight Rises \n",1723"4 49529 John Carter \n",1724"\n",1725" overview \\\n",1726"0 In the 22nd century, a paraplegic Marine is di... \n",1727"1 Captain Barbossa, long believed to be dead, ha... \n",1728"2 A cryptic message from Bond’s past sends him o... \n",1729"3 Following the death of District Attorney Harve... \n",1730"4 John Carter is a war-weary, former military ca... \n",1731"\n",1732" genres \\\n",1733"0 [Action, Adventure, Fantasy, ScienceFiction] \n",1734"1 [Adventure, Fantasy, Action] \n",1735"2 [Action, Adventure, Crime] \n",1736"3 [Action, Crime, Drama, Thriller] \n",1737"4 [Action, Adventure, ScienceFiction] \n",1738"\n",1739" keywords \\\n",1740"0 [cultureclash, future, spacewar, spacecolony, ... \n",1741"1 [ocean, drugabuse, exoticisland, eastindiatrad... \n",1742"2 [spy, basedonnovel, secretagent, sequel, mi6, ... \n",1743"3 [dccomics, crimefighter, terrorist, secretiden... \n",1744"4 [basedonnovel, mars, medallion, spacetravel, p... \n",1745"\n",1746" cast crew \n",1747"0 [SamWorthington, ZoeSaldana, SigourneyWeaver] [JamesCameron] \n",1748"1 [JohnnyDepp, OrlandoBloom, KeiraKnightley] [GoreVerbinski] \n",1749"2 [DanielCraig, ChristophWaltz, LéaSeydoux] [SamMendes] \n",1750"3 [ChristianBale, MichaelCaine, GaryOldman] [ChristopherNolan] \n",1751"4 [TaylorKitsch, LynnCollins, SamanthaMorton] [AndrewStanton] "1752]1753},1754"execution_count": 27,1755"metadata": {},1756"output_type": "execute_result"1757}1758],1759"source": [1760"movies.head()"1761]1762},1763{1764"cell_type": "code",1765"execution_count": 28,1766"id": "25b16cae",1767"metadata": {},1768"outputs": [],1769"source": [1770"movies['overview'] = movies['overview'].apply(lambda x:x.split())"1771]1772},1773{1774"cell_type": "code",1775"execution_count": 29,1776"id": "a4cf3ba0",1777"metadata": {},1778"outputs": [],1779"source": [1780"movies['tags'] = movies['overview'] + movies['genres'] + movies['keywords'] + movies['cast'] + movies['crew']"1781]1782},1783{1784"cell_type": "code",1785"execution_count": 30,1786"id": "5b85a9ee",1787"metadata": {},1788"outputs": [1789{1790"data": {1791"text/html": [1792"<div>\n",1793"<style scoped>\n",1794" .dataframe tbody tr th:only-of-type {\n",1795" vertical-align: middle;\n",1796" }\n",1797"\n",1798" .dataframe tbody tr th {\n",1799" vertical-align: top;\n",1800" }\n",1801"\n",1802" .dataframe thead th {\n",1803" text-align: right;\n",1804" }\n",1805"</style>\n",1806"<table border=\"1\" class=\"dataframe\">\n",1807" <thead>\n",1808" <tr style=\"text-align: right;\">\n",1809" <th></th>\n",1810" <th>movie_id</th>\n",1811" <th>title</th>\n",1812" <th>overview</th>\n",1813" <th>genres</th>\n",1814" <th>keywords</th>\n",1815" <th>cast</th>\n",1816" <th>crew</th>\n",1817" <th>tags</th>\n",1818" </tr>\n",1819" </thead>\n",1820" <tbody>\n",1821" <tr>\n",1822" <th>0</th>\n",1823" <td>19995</td>\n",1824" <td>Avatar</td>\n",1825" <td>[In, the, 22nd, century,, a, paraplegic, Marin...</td>\n",1826" <td>[Action, Adventure, Fantasy, ScienceFiction]</td>\n",1827" <td>[cultureclash, future, spacewar, spacecolony, ...</td>\n",1828" <td>[SamWorthington, ZoeSaldana, SigourneyWeaver]</td>\n",1829" <td>[JamesCameron]</td>\n",1830" <td>[In, the, 22nd, century,, a, paraplegic, Marin...</td>\n",1831" </tr>\n",1832" <tr>\n",1833" <th>1</th>\n",1834" <td>285</td>\n",1835" <td>Pirates of the Caribbean: At World's End</td>\n",1836" <td>[Captain, Barbossa,, long, believed, to, be, d...</td>\n",1837" <td>[Adventure, Fantasy, Action]</td>\n",1838" <td>[ocean, drugabuse, exoticisland, eastindiatrad...</td>\n",1839" <td>[JohnnyDepp, OrlandoBloom, KeiraKnightley]</td>\n",1840" <td>[GoreVerbinski]</td>\n",1841" <td>[Captain, Barbossa,, long, believed, to, be, d...</td>\n",1842" </tr>\n",1843" <tr>\n",1844" <th>2</th>\n",1845" <td>206647</td>\n",1846" <td>Spectre</td>\n",1847" <td>[A, cryptic, message, from, Bond’s, past, send...</td>\n",1848" <td>[Action, Adventure, Crime]</td>\n",1849" <td>[spy, basedonnovel, secretagent, sequel, mi6, ...</td>\n",1850" <td>[DanielCraig, ChristophWaltz, LéaSeydoux]</td>\n",1851" <td>[SamMendes]</td>\n",1852" <td>[A, cryptic, message, from, Bond’s, past, send...</td>\n",1853" </tr>\n",1854" <tr>\n",1855" <th>3</th>\n",1856" <td>49026</td>\n",1857" <td>The Dark Knight Rises</td>\n",1858" <td>[Following, the, death, of, District, Attorney...</td>\n",1859" <td>[Action, Crime, Drama, Thriller]</td>\n",1860" <td>[dccomics, crimefighter, terrorist, secretiden...</td>\n",1861" <td>[ChristianBale, MichaelCaine, GaryOldman]</td>\n",1862" <td>[ChristopherNolan]</td>\n",1863" <td>[Following, the, death, of, District, Attorney...</td>\n",1864" </tr>\n",1865" <tr>\n",1866" <th>4</th>\n",1867" <td>49529</td>\n",1868" <td>John Carter</td>\n",1869" <td>[John, Carter, is, a, war-weary,, former, mili...</td>\n",1870" <td>[Action, Adventure, ScienceFiction]</td>\n",1871" <td>[basedonnovel, mars, medallion, spacetravel, p...</td>\n",1872" <td>[TaylorKitsch, LynnCollins, SamanthaMorton]</td>\n",1873" <td>[AndrewStanton]</td>\n",1874" <td>[John, Carter, is, a, war-weary,, former, mili...</td>\n",1875" </tr>\n",1876" </tbody>\n",1877"</table>\n",1878"</div>"1879],1880"text/plain": [1881" movie_id title \\\n",1882"0 19995 Avatar \n",1883"1 285 Pirates of the Caribbean: At World's End \n",1884"2 206647 Spectre \n",1885"3 49026 The Dark Knight Rises \n",1886"4 49529 John Carter \n",1887"\n",1888" overview \\\n",1889"0 [In, the, 22nd, century,, a, paraplegic, Marin... \n",1890"1 [Captain, Barbossa,, long, believed, to, be, d... \n",1891"2 [A, cryptic, message, from, Bond’s, past, send... \n",1892"3 [Following, the, death, of, District, Attorney... \n",1893"4 [John, Carter, is, a, war-weary,, former, mili... \n",1894"\n",1895" genres \\\n",1896"0 [Action, Adventure, Fantasy, ScienceFiction] \n",1897"1 [Adventure, Fantasy, Action] \n",1898"2 [Action, Adventure, Crime] \n",1899"3 [Action, Crime, Drama, Thriller] \n",1900"4 [Action, Adventure, ScienceFiction] \n",1901"\n",1902" keywords \\\n",1903"0 [cultureclash, future, spacewar, spacecolony, ... \n",1904"1 [ocean, drugabuse, exoticisland, eastindiatrad... \n",1905"2 [spy, basedonnovel, secretagent, sequel, mi6, ... \n",1906"3 [dccomics, crimefighter, terrorist, secretiden... \n",1907"4 [basedonnovel, mars, medallion, spacetravel, p... \n",1908"\n",1909" cast crew \\\n",1910"0 [SamWorthington, ZoeSaldana, SigourneyWeaver] [JamesCameron] \n",1911"1 [JohnnyDepp, OrlandoBloom, KeiraKnightley] [GoreVerbinski] \n",1912"2 [DanielCraig, ChristophWaltz, LéaSeydoux] [SamMendes] \n",1913"3 [ChristianBale, MichaelCaine, GaryOldman] [ChristopherNolan] \n",1914"4 [TaylorKitsch, LynnCollins, SamanthaMorton] [AndrewStanton] \n",1915"\n",1916" tags \n",1917"0 [In, the, 22nd, century,, a, paraplegic, Marin... \n",1918"1 [Captain, Barbossa,, long, believed, to, be, d... \n",1919"2 [A, cryptic, message, from, Bond’s, past, send... \n",1920"3 [Following, the, death, of, District, Attorney... \n",1921"4 [John, Carter, is, a, war-weary,, former, mili... "1922]1923},1924"execution_count": 30,1925"metadata": {},1926"output_type": "execute_result"1927}1928],1929"source": [1930"movies.head()"1931]1932},1933{1934"cell_type": "code",1935"execution_count": 31,1936"id": "c8e5db66",1937"metadata": {},1938"outputs": [],1939"source": [1940"new = movies.drop(columns=['overview','genres','keywords','cast','crew'])\n",1941"#new.head()"1942]1943},1944{1945"cell_type": "code",1946"execution_count": 32,1947"id": "86337c80",1948"metadata": {},1949"outputs": [1950{1951"data": {1952"text/html": [1953"<div>\n",1954"<style scoped>\n",1955" .dataframe tbody tr th:only-of-type {\n",1956" vertical-align: middle;\n",1957" }\n",1958"\n",1959" .dataframe tbody tr th {\n",1960" vertical-align: top;\n",1961" }\n",1962"\n",1963" .dataframe thead th {\n",1964" text-align: right;\n",1965" }\n",1966"</style>\n",1967"<table border=\"1\" class=\"dataframe\">\n",1968" <thead>\n",1969" <tr style=\"text-align: right;\">\n",1970" <th></th>\n",1971" <th>movie_id</th>\n",1972" <th>title</th>\n",1973" <th>tags</th>\n",1974" </tr>\n",1975" </thead>\n",1976" <tbody>\n",1977" <tr>\n",1978" <th>0</th>\n",1979" <td>19995</td>\n",1980" <td>Avatar</td>\n",1981" <td>In the 22nd century, a paraplegic Marine is di...</td>\n",1982" </tr>\n",1983" <tr>\n",1984" <th>1</th>\n",1985" <td>285</td>\n",1986" <td>Pirates of the Caribbean: At World's End</td>\n",1987" <td>Captain Barbossa, long believed to be dead, ha...</td>\n",1988" </tr>\n",1989" <tr>\n",1990" <th>2</th>\n",1991" <td>206647</td>\n",1992" <td>Spectre</td>\n",1993" <td>A cryptic message from Bond’s past sends him o...</td>\n",1994" </tr>\n",1995" <tr>\n",1996" <th>3</th>\n",1997" <td>49026</td>\n",1998" <td>The Dark Knight Rises</td>\n",1999" <td>Following the death of District Attorney Harve...</td>\n",2000" </tr>\n",2001" <tr>\n",2002" <th>4</th>\n",2003" <td>49529</td>\n",2004" <td>John Carter</td>\n",2005" <td>John Carter is a war-weary, former military ca...</td>\n",2006" </tr>\n",2007" </tbody>\n",2008"</table>\n",2009"</div>"2010],2011"text/plain": [2012" movie_id title \\\n",2013"0 19995 Avatar \n",2014"1 285 Pirates of the Caribbean: At World's End \n",2015"2 206647 Spectre \n",2016"3 49026 The Dark Knight Rises \n",2017"4 49529 John Carter \n",2018"\n",2019" tags \n",2020"0 In the 22nd century, a paraplegic Marine is di... \n",2021"1 Captain Barbossa, long believed to be dead, ha... \n",2022"2 A cryptic message from Bond’s past sends him o... \n",2023"3 Following the death of District Attorney Harve... \n",2024"4 John Carter is a war-weary, former military ca... "2025]2026},2027"execution_count": 32,2028"metadata": {},2029"output_type": "execute_result"2030}2031],2032"source": [2033"new['tags'] = new['tags'].apply(lambda x: \" \".join(x))\n",2034"new.head()"2035]2036},2037{2038"cell_type": "code",2039"execution_count": 33,2040"id": "b2096072",2041"metadata": {},2042"outputs": [],2043"source": [2044"from sklearn.feature_extraction.text import CountVectorizer\n",2045"cv = CountVectorizer(max_features=5000,stop_words='english')"2046]2047},2048{2049"cell_type": "code",2050"execution_count": 34,2051"id": "78653bfc",2052"metadata": {},2053"outputs": [],2054"source": [2055"vector = cv.fit_transform(new['tags']).toarray()"2056]2057},2058{2059"cell_type": "code",2060"execution_count": 35,2061"id": "90f36092",2062"metadata": {},2063"outputs": [2064{2065"data": {2066"text/plain": [2067"array([[0, 0, 0, ..., 0, 0, 0],\n",2068" [0, 0, 0, ..., 0, 0, 0],\n",2069" [0, 0, 0, ..., 0, 0, 0],\n",2070" ...,\n",2071" [0, 0, 0, ..., 0, 0, 0],\n",2072" [0, 0, 0, ..., 0, 0, 0],\n",2073" [0, 0, 0, ..., 0, 0, 0]], dtype=int64)"2074]2075},2076"execution_count": 35,2077"metadata": {},2078"output_type": "execute_result"2079}2080],2081"source": [2082"vector"2083]2084},2085{2086"cell_type": "code",2087"execution_count": 36,2088"id": "e4a28860",2089"metadata": {},2090"outputs": [2091{2092"data": {2093"text/plain": [2094"array([0, 0, 0, ..., 0, 0, 0], dtype=int64)"2095]2096},2097"execution_count": 36,2098"metadata": {},2099"output_type": "execute_result"2100}2101],2102"source": [2103"vector[0]"2104]2105},2106{2107"cell_type": "code",2108"execution_count": 37,2109"id": "d7809e16",2110"metadata": {},2111"outputs": [],2112"source": [2113"# cv.get_feature_names()"2114]2115},2116{2117"cell_type": "code",2118"execution_count": 38,2119"id": "4b71c2f2",2120"metadata": {},2121"outputs": [2122{2123"data": {2124"text/plain": [2125"(4806, 5000)"2126]2127},2128"execution_count": 38,2129"metadata": {},2130"output_type": "execute_result"2131}2132],2133"source": [2134"vector.shape"2135]2136},2137{2138"cell_type": "code",2139"execution_count": 39,2140"id": "12a944fa",2141"metadata": {},2142"outputs": [],2143"source": [2144"import nltk"2145]2146},2147{2148"cell_type": "code",2149"execution_count": 40,2150"id": "4abe934e",2151"metadata": {},2152"outputs": [],2153"source": [2154"from nltk.stem.porter import PorterStemmer\n",2155"ps= PorterStemmer()"2156]2157},2158{2159"cell_type": "code",2160"execution_count": 41,2161"id": "0ff40f63",2162"metadata": {},2163"outputs": [],2164"source": [2165"def stem(text):\n",2166" y=[]\n",2167" \n",2168" for i in text.split():\n",2169" y.append(ps.stem(i))\n",2170" \n",2171" return \" \".join(y)"2172]2173},2174{2175"cell_type": "code",2176"execution_count": 42,2177"id": "b98e0748",2178"metadata": {},2179"outputs": [],2180"source": [2181"new['tags'] = new['tags'].apply(stem) #stem is used to combine words of same meaning in one word"2182]2183},2184{2185"cell_type": "code",2186"execution_count": 43,2187"id": "a8bab646",2188"metadata": {},2189"outputs": [],2190"source": [2191"from sklearn.metrics.pairwise import cosine_similarity \n"2192]2193},2194{2195"cell_type": "code",2196"execution_count": 44,2197"id": "03d53db1",2198"metadata": {},2199"outputs": [],2200"source": [2201"similarity = cosine_similarity(vector) #calculate angle btw to vectors and find the sisimilarities\n"2202]2203},2204{2205"cell_type": "code",2206"execution_count": 45,2207"id": "289f5e53",2208"metadata": {},2209"outputs": [2210{2211"data": {2212"text/plain": [2213"array([[1. , 0.08964215, 0.06071767, ..., 0.02519763, 0.0277885 ,\n",2214" 0. ],\n",2215" [0.08964215, 1. , 0.06350006, ..., 0.02635231, 0. ,\n",2216" 0. ],\n",2217" [0.06071767, 0.06350006, 1. , ..., 0.02677398, 0. ,\n",2218" 0. ],\n",2219" ...,\n",2220" [0.02519763, 0.02635231, 0.02677398, ..., 1. , 0.07352146,\n",2221" 0.04774099],\n",2222" [0.0277885 , 0. , 0. , ..., 0.07352146, 1. ,\n",2223" 0.05264981],\n",2224" [0. , 0. , 0. , ..., 0.04774099, 0.05264981,\n",2225" 1. ]])"2226]2227},2228"execution_count": 45,2229"metadata": {},2230"output_type": "execute_result"2231}2232],2233"source": [2234"similarity\n"2235]2236},2237{2238"cell_type": "code",2239"execution_count": 46,2240"id": "878adeff",2241"metadata": {},2242"outputs": [2243{2244"data": {2245"text/plain": [2246"744"2247]2248},2249"execution_count": 46,2250"metadata": {},2251"output_type": "execute_result"2252}2253],2254"source": [2255"new[new['title'] == 'The Lego Movie'].index[0]\n"2256]2257},2258{2259"cell_type": "code",2260"execution_count": 47,2261"id": "00337383",2262"metadata": {},2263"outputs": [],2264"source": [2265"def recommend(movie):\n",2266" index = new[new['title'] == movie].index[0]\n",2267" movie_list = sorted(list(enumerate(similarity[index])),reverse=True,key = lambda x: x[1])\n",2268" for i in movie_list[1:6]:\n",2269" print(new.iloc[i[0]].title)"2270]2271},2272{2273"cell_type": "code",2274"execution_count": 48,2275"id": "67047f9b",2276"metadata": {},2277"outputs": [2278{2279"name": "stdout",2280"output_type": "stream",2281"text": [2282"The Dark Knight\n",2283"The Dark Knight Rises\n",2284"Batman\n",2285"Batman & Robin\n",2286"Batman\n"2287]2288}2289],2290"source": [2291"recommend('Batman Begins')\n"2292]2293},2294{2295"cell_type": "code",2296"execution_count": 49,2297"id": "ef4ae40b",2298"metadata": {},2299"outputs": [],2300"source": [2301"import pickle"2302]2303},2304{2305"cell_type": "code",2306"execution_count": 50,2307"id": "421e14ba",2308"metadata": {},2309"outputs": [],2310"source": [2311"pickle.dump(new,open('movie_list.pkl','wb'))\n",2312"pickle.dump(similarity,open('similarity.pkl','wb'))"2313]2314},2315{2316"cell_type": "code",2317"execution_count": 53,2318"id": "e8c97e90",2319"metadata": {},2320"outputs": [2321{2322"name": "stdout",2323"output_type": "stream",2324"text": [2325"Requirement already satisfied: pandas in d:\\desktop\\anaconda23\\lib\\site-packages (1.5.3)\n",2326"Requirement already satisfied: numpy>=1.21.0 in d:\\desktop\\anaconda23\\lib\\site-packages (from pandas) (1.23.5)\n",2327"Requirement already satisfied: python-dateutil>=2.8.1 in d:\\desktop\\anaconda23\\lib\\site-packages (from pandas) (2.8.2)\n",2328"Requirement already satisfied: pytz>=2020.1 in d:\\desktop\\anaconda23\\lib\\site-packages (from pandas) (2022.7)\n",2329"Requirement already satisfied: six>=1.5 in d:\\desktop\\anaconda23\\lib\\site-packages (from python-dateutil>=2.8.1->pandas) (1.16.0)\n"2330]2331}2332],2333"source": [2334"!pip install pandas --version"2335]2336},2337{2338"cell_type": "code",2339"execution_count": 55,2340"id": "33b2c364",2341"metadata": {},2342"outputs": [2343{2344"name": "stdout",2345"output_type": "stream",2346"text": [2347"1.5.3\n"2348]2349}2350],2351"source": [2352"!python -c \"import pandas as pd; print(pd.__version__)\"\n"2353]2354}2355],2356"metadata": {2357"kernelspec": {2358"display_name": "Python 3 (ipykernel)",2359"language": "python",2360"name": "python3"2361},2362"language_info": {2363"codemirror_mode": {2364"name": "ipython",2365"version": 32366},2367"file_extension": ".py",2368"mimetype": "text/x-python",2369"name": "python",2370"nbconvert_exporter": "python",2371"pygments_lexer": "ipython3",2372"version": "3.10.9"2373}2374},2375"nbformat": 4,2376"nbformat_minor": 52377}2378