ium_151636/zadanie1.ipynb

import pandas as pd
movies = pd.read_csv("25k_IMDb_movie_Dataset.csv")
movies

	movie title	Run Time	Rating	User Rating	Generes	Overview	Plot Kyeword	Director	Top 5 Casts	Writer	year	path
0	Top Gun: Maverick	$170,000,000 (estimated)	8.6	187K	['Action', 'Drama']	After more than thirty years of service as one...	['fighter jet', 'sequel', 'u.s. navy', 'fighte...	Joseph Kosinski	['Jack Epps Jr.', 'Peter Craig', 'Tom Cruise',...	Jim Cash	-2022	/title/tt1745960/
1	Jurassic World Dominion	2 hours 27 minutes	6	56K	['Action', 'Adventure', 'Sci-Fi']	Four years after the destruction of Isla Nubla...	['dinosaur', 'jurassic park', 'tyrannosaurus r...	Colin Trevorrow	['Colin Trevorrow', 'Derek Connolly', 'Chris P...	Emily Carmichael	-2022	/title/tt8041270/
2	Top Gun	$15,000,000 (estimated)	6.9	380K	['Action', 'Drama']	As students at the United States Navy's elite ...	['pilot', 'male camaraderie', 'u.s. navy', 'gr...	Tony Scott	['Jack Epps Jr.', 'Ehud Yonay', 'Tom Cruise', ...	Jim Cash	-1986	/title/tt0092099/
3	Lightyear	$71,101,257	5.2	32K	['Animation', 'Action', 'Adventure']	While spending years attempting to return home...	['galaxy', 'spaceship', 'robot', 'rocket', 'sp...	Angus MacLane	['Jason Headley', 'Matthew Aldrich', 'Chris Ev...	Angus MacLane	-2022	/title/tt10298810/
4	Spiderhead	not-released	5.4	23K	['Action', 'Crime', 'Drama']	In the near future, convicts are offered the c...	['discover', 'medical', 'test', 'reality', 'fi...	Joseph Kosinski	['Rhett Reese', 'Paul Wernick', 'Chris Hemswor...	George Saunders	-2022	/title/tt9783600/
...	...	...	...	...	...	...	...	...	...	...	...	...
24397	Delicatessen	FRF 24,000,000 (estimated)	7.6	85K	['Comedy', 'Crime']	Post-apocalyptic surrealist black comedy about...	['surrealist', 'black comedy', 'human meat', '...	Marc Caro	['Jean-Pierre Jeunet', 'Marc Caro', 'Gilles Ad...	Jean-Pierre Jeunet	-1991	/title/tt0101700/
24398	Bitch Ass	not-released	5.5	52	['Crime', 'Horror']	A gang initiation goes wrong when a group of f...	[]	Bill Posley	['Bill Posley', 'Teon Kelley', 'Tunde Laleye',...	Jonathan Colomb	-2022	/title/tt13991504/
24399	Bullwhip	not-released	5.1	398	['Crime', 'Romance', 'Western']	In order to avoid the hangman's noose, a cowbo...	['taming of the shrew', 'fur trader', 'busines...	Harmon Jones	['Guy Madison', 'Rhonda Fleming', 'James Griff...	Adele Buffington	-1958	/title/tt0051438/
24400	The Freshman	1 hour 42 minutes	6.4	20K	['Comedy', 'Crime']	An N.Y.C. film school student accepts a job wi...	['endangered species', 'fish out of water', 'g...	Andrew Bergman	['Marlon Brando', 'Matthew Broderick', 'Bruno ...	Andrew Bergman	-1990	/title/tt0099615/
24401	Guys and Dolls	$5,500,000 (estimated)	7.1	18K	['Comedy', 'Crime', 'Musical']	In New York, a gambler is challenged to take a...	['mission', 'gambler', 'new york city', 'based...	Joseph L. Mankiewicz	['Abe Burrows', 'Damon Runyon', 'Marlon Brando...	Jo Swerling	-1955	/title/tt0048140/

24402 rows × 12 columns

import numpy as np

train, dev, test = np.split(movies.sample(frac=1, random_state=69), [int(.6*len(movies)), int(.8*len(movies))])

train

	movie title	Run Time	Rating	User Rating	Generes	Overview	Plot Kyeword	Director	Top 5 Casts	Writer	year	path
4782	Golmaal: Fun Unlimited	2 hours 30 minutes	7.4	17K	['Action', 'Comedy', 'Drama']	Four runaway crooks take shelter in a bungalow...	['blind couple', 'friend', 'mute', 'slacker', ...	Rohit Shetty	['Ajay Devgn', 'Arshad Warsi', 'Sharman Joshi'...	Neeraj Vora	-2006	/title/tt0495034/
9291	The Remains of the Day	2 hours 14 minutes	7.8	74K	['Drama', 'Romance']	A butler who sacrificed body and soul to servi...	['class differences', 'butler', 'housekeeper',...	James Ivory	['Ruth Prawer Jhabvala', 'Anthony Hopkins', 'E...	Kazuo Ishiguro	-1993	/title/tt0107943/
3137	Blue Iguana	1 hour 40 minutes	5.6	5.2K	['Action', 'Comedy', 'Crime']	Ex-jailbirds Eddie and Paul are on parole and ...	['singing in a car', 'reference to the world c...	Hadi Hajaig	['Sam Rockwell', 'Phoebe Fox', 'Ben Schwartz',...	Hadi Hajaig	-2018	/title/tt2316479/
17052	Nocturna	1 hour 20 minutes	7.2	2.4K	['Animation', 'Adventure', 'Family']	An orphan boy named Tim is afraid of the dark....	['orphan', 'night', 'cat', 'one word title', '...	Adrià García	['Adrià García', 'Víctor Maldonado', 'Teresa V...	Víctor Maldonado	-2007	/title/tt0836682/
675	Alexander	$155,000,000 (estimated)	5.6	169K	['Action', 'Biography', 'Drama']	Alexander, the King of Macedonia and one of th...	['ancient greece', 'greek', 'macedonia', 'sex ...	Oliver Stone	['Christopher Kyle', 'Laeta Kalogridis', 'Coli...	Oliver Stone	-2004	/title/tt0346491/
...	...	...	...	...	...	...	...	...	...	...	...	...
6365	Yalghaar	2 hours 38 minutes	6.4	1K	['Action', 'Romance', 'War']	The film "explores what happens in the lives o...	['live']	Hassan Rana	['Shaan Shahid', 'Humayun Saeed', 'Adnan Siddi...	Hassan Rana	-2017	/title/tt3945864/
7172	Brown of Harvard	not-released	6.2	1.5K	['Action', 'Drama', 'Romance']	Tom Brown shows up at Harvard, confident and a...	['harvard', 'no homoeroticism', 'pre code film...	Jack Conway	['Donald Ogden Stewart', 'Andrew Percival Youn...	Rida Johnson Young	-1926	/title/tt0016690/
12644	Deseo	1 hour 37 minutes	4.6	417	['Comedy', 'Drama', 'Romance']	A succession of erotic encounters weaved into ...	['smoking marijuana', 'female full frontal nud...	Antonio Zavala Kugler	['Antonio Zavala Kugler', 'Christian Bach', 'A...	Arthur Schnitzler	-2013	/title/tt1236434/
23970	Saezuru Tori Wa Habatakanai: The Clouds Gather	1 hour 25 minutes	6.9	681	['Animation', 'Crime', 'Drama']	Yashiro is the president of the Shinseikai Ent...	['yaoi', 'boys love', 'gay', 'anime', 'yakuza'...	Kaori Makita	['Kou Yoneda', 'Tarusuke Shingaki', 'Wataru Ha...	Hiroshi Seko	-2020	/title/tt10675392/
13423	99	2 hours 15 minutes	7.3	2.9K	['Comedy', 'Crime', 'Drama']	A gangster deputes two of his men to recover m...	['cricket the sport', 'briefcase', 'caper', 'e...	Krishna D.K.	['Raj Nidimoru', 'Krishna D.K.', 'Sita Menon',...	Raj Nidimoru	(I) (2009)	/title/tt1370429/

14641 rows × 12 columns

dev

	movie title	Run Time	Rating	User Rating	Generes	Overview	Plot Kyeword	Director	Top 5 Casts	Writer	year	path
20683	The Animals	not-released	4.2	398	['Western']	A woman tracks down the five men who raped her...	['rape and revenge', 'arizona territory', 'ari...	Ron Joy	['Michele Carey', 'Henry Silva', 'Keenan Wynn'...	Richard Bakalyan	-1970	/title/tt0065407/
8040	The Lincoln Lawyer	$40,000,000 (estimated)	7.3	236K	['Crime', 'Drama', 'Mystery']	A lawyer defending a wealthy man begins to bel...	['defense lawyer', 'plot twist', 'drug rehabil...	Brad Furman	['Michael Connelly', 'Matthew McConaughey', 'M...	John Romano	-2011	/title/tt1189340/
23664	What's the Worst That Could Happen?	1 hour 34 minutes	5.4	16K	['Comedy', 'Crime']	A rich man catches a thief burglarizing his ho...	['breaking and entering', 'bankruptcy', 'quest...	Sam Weisman	['Matthew Chapman', 'Martin Lawrence', 'Danny ...	Donald E. Westlake	-2001	/title/tt0161083/
9078	Cleopatra	not-released	6.2	652	['Animation', 'History', 'Romance']	In order to foil the enemy aliens' "Cleopatra ...	['adult anime', 'adult animation', 'anime', 'd...	Osamu Tezuka	['Osamu Tezuka', 'Shigemi Satoyoshi', 'Chinats...	Eiichi Yamamoto	-1963	/title/tt0056937/
17118	Seven Cities of Gold	1 hour 43 minutes	5.9	479	['Adventure', 'Biography', 'History']	In 1769, a Spanish expedition to California se...	['limping man', 'prologue', 'voice over narrat...	Robert D. Webb	['John C. Higgins', 'Joseph Petracca', 'Richar...	Richard L. Breen	-1955	/title/tt0048603/
...	...	...	...	...	...	...	...	...	...	...	...	...
1248	The Wall	1 hour 28 minutes	6.2	27K	['Action', 'Drama', 'Thriller']	Two American Soldiers are trapped by a lethal ...	['sniper', 'soldier', 'deception', 'wound', 'd...	Doug Liman	['Aaron Taylor-Johnson', 'John Cena', 'Laith N...	Dwain Worrell	(II) (2017)	/title/tt4218696/
7091	Commando 3	2 hours 13 minutes	5.7	3K	['Action', 'Adventure', 'Thriller']	Karan goes to London to stop a terrorist attac...	['chase']	Aditya Datt	['Junaid Wasi', 'Gulshan Devaiah', 'Robin Chau...	Darius Yarmil	-2019	/title/tt8983168/
15635	Adam Resurrected	1 hour 46 minutes	6.2	4.2K	['Drama', 'War']	In the aftermath of World War II, a former cir...	['man wears a white suit', 'desert hotel', 'ti...	Paul Schrader	['Noah Stollman', 'Jeff Goldblum', 'Willem Daf...	Yoram Kaniuk	-2008	/title/tt0479341/
12914	Main Street	1 hour 32 minutes	4.7	2.9K	['Drama']	Durham is slowly dying like the tobacco busine...	['economic depression', 'reference to lucky st...	John Doyle	['Colin Firth', 'Ellen Burstyn', 'Patricia Cla...	Horton Foote	-2010	/title/tt1365483/
2825	The Outrage	1 hour 36 minutes	6.2	2.2K	['Crime', 'Drama', 'Western']	Travelers in the 1870s Southwest discuss a rec...	['highwayman', 'man bound and gagged', 'gun du...	Martin Ritt	['Akira Kurosawa', 'Ryûnosuke Akutagawa', 'Pau...	Michael Kanin	-1964	/title/tt0058437/

4880 rows × 12 columns

test

	movie title	Run Time	Rating	User Rating	Generes	Overview	Plot Kyeword	Director	Top 5 Casts	Writer	year	path
16788	High in the Clouds	not-released	no-rating	0	['Animation', 'Adventure', 'Comedy']	A squirrel embarks on a journey to find an ani...	['journey', 'cloud', 'animal', 'friend', 'trag...	Timothy Reckart	['Jon Croker', 'Geoff Dunbar', 'Timothy Reckar...	Philip Ardagh	-2022	/title/tt1458167/
17155	Curious George 3: Back to the Jungle	1 hour 21 minutes	5.4	598	['Animation', 'Adventure', 'Comedy']	Curious George goes on an epic adventure to sp...	['sequel', 'third part', 'jungle', 'curious ge...	Phil Weinstein	['H.A. Rey', 'Chuck Tately', 'Frank Welker', '...	Margret Rey	-2015	/title/tt4622340/
5985	Tarkan: Altin Madalyon	not-released	6.2	1.6K	['Action', 'Adventure', 'History']	The story of Tarkan and his friends efforts to...	['black magic', 'blood', 'axe', 'snake', 'pros...	Mehmet Aslan	['Sadik Sendil', 'Kartal Tibet', 'Eva Bender',...	Sezgin Burak	-1973	/title/tt0274933/
1640	Harley Davidson and the Marlboro Man	$23,000,000 (estimated)	6.1	20K	['Action', 'Crime', 'Drama']	Forced by the imminent foreclosure of their fr...	['swimming pool', 'night club', 'voyeurism', '...	Simon Wincer	['Mickey Rourke', 'Don Johnson', 'Chelsea Fiel...	Don Michael Paul	-1991	/title/tt0102005/
9826	I Am All Girls	not-released	5.9	5.8K	['Crime', 'Drama', 'Mystery']	A special crimes investigator forms an unlikel...	['child', 'bond', 'investigator', 'murder', 'c...	Donovan Marsh	['Marcell Greeff', 'Emile Leuvennink', 'Erica ...	Wayne Fitzjohn	-2021	/title/tt9013182/
...	...	...	...	...	...	...	...	...	...	...	...	...
11825	The Apostle	2 hours 14 minutes	7.2	14K	['Drama']	After his happy life spins out of control, a p...	['southern gothic', 'timeframe 1930s', 'preach...	Robert Duvall	['Robert Duvall', 'Todd Allen', 'Paul Bagget',...	Robert Duvall	-1997	/title/tt0118632/
14740	My Normal	not-released	4.1	756	['Drama', 'Romance']	A lesbian dominatrix finds a way to use her un...	['dominatrix', 'lesbian', 'lesbian kiss', 'blo...	Irving Schwartz	['Renee Garzon', 'Keith Planit', 'Nicole LaLib...	Abdul Malik Abbott	-2009	/title/tt1117983/
9818	Lust och fägring stor	2 hours 10 minutes	6.9	6.5K	['Drama', 'Romance', 'War']	Malmö, Sweden during the Second World War. Sti...	['extramarital affair', 'teacher student sex',...	Bo Widerberg	['Johan Widerberg', 'Marika Lagercrantz', 'Tom...	Bo Widerberg	-1995	/title/tt0113720/
4041	RoboCop Returns	not-released	no-rating	0	['Action', 'Adventure', 'Crime']	RoboCop returns to fight crime in Detroit.	['sequel', 'reboot', 'non comic book superhero...	Abe Forsythe	['Edward Neumeier', 'Justin Rhodes', 'Abe Fors...	Michael Miner	NaN	/title/tt8688550/
23755	The Maiden Heist	1 hour 30 minutes	6	17K	['Comedy', 'Crime']	A comedy centered on three museum security gua...	['heist crime', 'caper crime', 'forgery', 'hei...	Peter Hewitt	['Christopher Walken', 'Joseph McKenna', 'Wynn...	Michael LeSieur	-2009	/title/tt1107860/

4881 rows × 12 columns

movies.describe(include='all')

	movie title	Run Time	Rating	User Rating	Generes	Overview	Plot Kyeword	Director	Top 5 Casts	Writer	year	path
count	24402	24402	24402	24402	24402	24158	24402	24402	24402	24402	23624	24402
unique	23922	1556	91	1684	746	23957	21546	11604	24211	15562	250	23922
top	Rage	not-released	no-rating	0	['Drama']	none	[]	See company contact information	['See producer', 'See preliminary cast']	See writer	-2022	/title/tt0114224/
freq	4	8475	1740	1740	943	142	1696	142	142	142	1201	4

movies["movie title"].value_counts()

Rage                                         4
The Killer                                   4
The Beast                                    4
Spiral                                       4
The Silence                                  3
                                            ..
The Mule                                     1
Donnie Brasco                                1
Little Miss Sunshine                         1
Three Billboards Outside Ebbing, Missouri    1
Guys and Dolls                               1
Name: movie title, Length: 23922, dtype: int64

movies["Run Time"].value_counts()

not-released                  8475
1 hour 30 minutes              503
1 hour 35 minutes              376
1 hour 38 minutes              350
1 hour 31 minutes              338
                              ... 
$14,492                          1
$181,415                         1
$11,060,485                      1
$1,043,910                       1
FRF 24,000,000 (estimated)       1
Name: Run Time, Length: 1556, dtype: int64

movies["Rating"].value_counts()

no-rating    1740
6.4           852
6.2           847
6.1           819
6.3           809
             ... 
9.9             2
9.8             2
9.4             2
1               2
9.5             2
Name: Rating, Length: 91, dtype: int64

movies["User Rating"].value_counts()

0       1740
11K      325
1.2K     323
1.1K     315
1.3K     295
        ... 
501K       1
769K       1
321K       1
991K       1
347K       1
Name: User Rating, Length: 1684, dtype: int64

movies["Generes"].value_counts()

['Drama']                            943
['Action', 'Crime', 'Drama']         867
['Crime', 'Drama', 'Thriller']       609
['Comedy', 'Drama', 'Romance']       608
['Crime', 'Drama']                   550
                                    ... 
['Drama', 'Romance', 'Crime']          1
['Drama', 'Crime', 'Mystery']          1
['Family', 'Adventure', 'Comedy']      1
['Crime', 'Mystery', 'Horror']         1
['Crime', 'Romance', 'Western']        1
Name: Generes, Length: 746, dtype: int64