In [2]:
import os
import pandas as pd
import settings
import etl
%matplotlib inline
%load_ext watermark
%watermark -d -t -v -m -p pea,pandas
2017-07-03 16:52:06
CPython 3.6.1
IPython 6.1.0
pea 0.0.7
pandas 0.20.2
compiler : MSC v.1900 64 bit (AMD64)
system : Windows
release : 7
machine : AMD64
processor : Intel64 Family 6 Model 42 Stepping 7, GenuineIntel
CPU cores : 8
interpreter: 64bit
In [3]:
data = etl.Data()
In [4]:
data.movie
Out[4]:
color
director_name
num_critic_for_reviews
duration
director_facebook_likes
actor_3_facebook_likes
actor_2_name
actor_1_facebook_likes
gross
genres
...
country
content_rating
budget
title_year
actor_2_facebook_likes
imdb_score
aspect_ratio
movie_facebook_likes
net
profitable
0
Color
James Cameron
723.0
178.0
0.0
855.0
Joel David Moore
1000.0
760505847.0
Action|Adventure|Fantasy|Sci-Fi
...
USA
PG-13
237000000.0
2009-01-01
936.0
7.9
1.78
33000
523505847.0
1
1
Color
Gore Verbinski
302.0
169.0
563.0
1000.0
Orlando Bloom
40000.0
309404152.0
Action|Adventure|Fantasy
...
USA
PG-13
300000000.0
2007-01-01
5000.0
7.1
2.35
0
9404152.0
1
2
Color
Sam Mendes
602.0
148.0
0.0
161.0
Rory Kinnear
11000.0
200074175.0
Action|Adventure|Thriller
...
UK
PG-13
245000000.0
2015-01-01
393.0
6.8
2.35
85000
-44925825.0
0
3
Color
Christopher Nolan
813.0
164.0
22000.0
23000.0
Christian Bale
27000.0
448130642.0
Action|Thriller
...
USA
PG-13
250000000.0
2012-01-01
23000.0
8.5
2.35
164000
198130642.0
1
4
NaN
Doug Walker
NaN
NaN
131.0
NaN
Rob Walker
131.0
NaN
Documentary
...
NaN
NaN
NaN
NaT
12.0
7.1
NaN
0
NaN
0
5
Color
Andrew Stanton
462.0
132.0
475.0
530.0
Samantha Morton
640.0
73058679.0
Action|Adventure|Sci-Fi
...
USA
PG-13
263700000.0
2012-01-01
632.0
6.6
2.35
24000
-190641321.0
0
6
Color
Sam Raimi
392.0
156.0
0.0
4000.0
James Franco
24000.0
336530303.0
Action|Adventure|Romance
...
USA
PG-13
258000000.0
2007-01-01
11000.0
6.2
2.35
0
78530303.0
1
7
Color
Nathan Greno
324.0
100.0
15.0
284.0
Donna Murphy
799.0
200807262.0
Adventure|Animation|Comedy|Family|Fantasy|Musi...
...
USA
PG
260000000.0
2010-01-01
553.0
7.8
1.85
29000
-59192738.0
0
8
Color
Joss Whedon
635.0
141.0
0.0
19000.0
Robert Downey Jr.
26000.0
458991599.0
Action|Adventure|Sci-Fi
...
USA
PG-13
250000000.0
2015-01-01
21000.0
7.5
2.35
118000
208991599.0
1
9
Color
David Yates
375.0
153.0
282.0
10000.0
Daniel Radcliffe
25000.0
301956980.0
Adventure|Family|Fantasy|Mystery
...
UK
PG
250000000.0
2009-01-01
11000.0
7.5
2.35
10000
51956980.0
1
10
Color
Zack Snyder
673.0
183.0
0.0
2000.0
Lauren Cohan
15000.0
330249062.0
Action|Adventure|Sci-Fi
...
USA
PG-13
250000000.0
2016-01-01
4000.0
6.9
2.35
197000
80249062.0
1
11
Color
Bryan Singer
434.0
169.0
0.0
903.0
Marlon Brando
18000.0
200069408.0
Action|Adventure|Sci-Fi
...
USA
PG-13
209000000.0
2006-01-01
10000.0
6.1
2.35
0
-8930592.0
0
12
Color
Marc Forster
403.0
106.0
395.0
393.0
Mathieu Amalric
451.0
168368427.0
Action|Adventure
...
UK
PG-13
200000000.0
2008-01-01
412.0
6.7
2.35
0
-31631573.0
0
13
Color
Gore Verbinski
313.0
151.0
563.0
1000.0
Orlando Bloom
40000.0
423032628.0
Action|Adventure|Fantasy
...
USA
PG-13
225000000.0
2006-01-01
5000.0
7.3
2.35
5000
198032628.0
1
14
Color
Gore Verbinski
450.0
150.0
563.0
1000.0
Ruth Wilson
40000.0
89289910.0
Action|Adventure|Western
...
USA
PG-13
215000000.0
2013-01-01
2000.0
6.5
2.35
48000
-125710090.0
0
15
Color
Zack Snyder
733.0
143.0
0.0
748.0
Christopher Meloni
15000.0
291021565.0
Action|Adventure|Fantasy|Sci-Fi
...
USA
PG-13
225000000.0
2013-01-01
3000.0
7.2
2.35
118000
66021565.0
1
16
Color
Andrew Adamson
258.0
150.0
80.0
201.0
Pierfrancesco Favino
22000.0
141614023.0
Action|Adventure|Family|Fantasy
...
USA
PG
225000000.0
2008-01-01
216.0
6.6
2.35
0
-83385977.0
0
17
Color
Joss Whedon
703.0
173.0
0.0
19000.0
Robert Downey Jr.
26000.0
623279547.0
Action|Adventure|Sci-Fi
...
USA
PG-13
220000000.0
2012-01-01
21000.0
8.1
1.85
123000
403279547.0
1
18
Color
Rob Marshall
448.0
136.0
252.0
1000.0
Sam Claflin
40000.0
241063875.0
Action|Adventure|Fantasy
...
USA
PG-13
250000000.0
2011-01-01
11000.0
6.7
2.35
58000
-8936125.0
0
19
Color
Barry Sonnenfeld
451.0
106.0
188.0
718.0
Michael Stuhlbarg
10000.0
179020854.0
Action|Adventure|Comedy|Family|Fantasy|Sci-Fi
...
USA
PG-13
225000000.0
2012-01-01
816.0
6.8
1.85
40000
-45979146.0
0
20
Color
Peter Jackson
422.0
164.0
0.0
773.0
Adam Brown
5000.0
255108370.0
Adventure|Fantasy
...
New Zealand
PG-13
250000000.0
2014-01-01
972.0
7.5
2.35
65000
5108370.0
1
21
Color
Marc Webb
599.0
153.0
464.0
963.0
Andrew Garfield
15000.0
262030663.0
Action|Adventure|Fantasy
...
USA
PG-13
230000000.0
2012-01-01
10000.0
7.0
2.35
56000
32030663.0
1
22
Color
Ridley Scott
343.0
156.0
0.0
738.0
William Hurt
891.0
105219735.0
Action|Adventure|Drama|History
...
USA
PG-13
200000000.0
2010-01-01
882.0
6.7
2.35
17000
-94780265.0
0
23
Color
Peter Jackson
509.0
186.0
0.0
773.0
Adam Brown
5000.0
258355354.0
Adventure|Fantasy
...
USA
PG-13
225000000.0
2013-01-01
972.0
7.9
2.35
83000
33355354.0
1
24
Color
Chris Weitz
251.0
113.0
129.0
1000.0
Eva Green
16000.0
70083519.0
Adventure|Family|Fantasy
...
USA
PG-13
180000000.0
2007-01-01
6000.0
6.1
2.35
0
-109916481.0
0
25
Color
Peter Jackson
446.0
201.0
0.0
84.0
Thomas Kretschmann
6000.0
218051260.0
Action|Adventure|Drama|Romance
...
New Zealand
PG-13
207000000.0
2005-01-01
919.0
7.2
2.35
0
11051260.0
1
26
Color
James Cameron
315.0
194.0
0.0
794.0
Kate Winslet
29000.0
658672302.0
Drama|Romance
...
USA
PG-13
200000000.0
1997-01-01
14000.0
7.7
2.35
26000
458672302.0
1
27
Color
Anthony Russo
516.0
147.0
94.0
11000.0
Scarlett Johansson
21000.0
407197282.0
Action|Adventure|Sci-Fi
...
USA
PG-13
250000000.0
2016-01-01
19000.0
8.2
2.35
72000
157197282.0
1
28
Color
Peter Berg
377.0
131.0
532.0
627.0
Alexander Skarsgård
14000.0
65173160.0
Action|Adventure|Sci-Fi|Thriller
...
USA
PG-13
209000000.0
2012-01-01
10000.0
5.9
2.35
44000
-143826840.0
0
29
Color
Colin Trevorrow
644.0
124.0
365.0
1000.0
Judy Greer
3000.0
652177271.0
Action|Adventure|Sci-Fi|Thriller
...
USA
PG-13
150000000.0
2015-01-01
2000.0
7.0
2.00
150000
502177271.0
1
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
5013
Color
Eric Eason
28.0
79.0
3.0
42.0
Panchito Gómez
93.0
NaN
Drama|Family
...
USA
NaN
24000.0
2002-01-01
46.0
7.0
1.78
61
NaN
0
5014
Color
Uwe Boll
58.0
80.0
892.0
492.0
Katharine Isabelle
986.0
NaN
Action|Crime|Thriller
...
Canada
R
NaN
2009-01-01
918.0
6.3
2.35
0
NaN
0
5015
Black and White
Richard Linklater
61.0
100.0
0.0
0.0
Richard Linklater
5.0
1227508.0
Comedy|Drama
...
USA
R
23000.0
1991-01-01
0.0
7.1
1.37
2000
1204508.0
1
5016
Color
Joseph Mazzella
NaN
90.0
0.0
9.0
Mikaal Bates
313.0
NaN
Crime|Drama|Thriller
...
USA
NaN
25000.0
2015-01-01
25.0
4.8
NaN
33
NaN
0
5017
Color
Travis Legge
1.0
90.0
138.0
138.0
Suzi Lorraine
370.0
NaN
Comedy|Romance
...
USA
NaN
22000.0
2013-01-01
184.0
3.3
1.78
200
NaN
0
5018
Color
Alex Kendrick
5.0
120.0
589.0
4.0
Lisa Arnold
51.0
NaN
Drama
...
USA
NaN
20000.0
2003-01-01
49.0
6.9
1.85
725
NaN
0
5019
Color
Marcus Nispel
43.0
91.0
158.0
265.0
Brittany Curran
630.0
NaN
Horror|Mystery|Thriller
...
USA
R
NaN
2015-01-01
512.0
4.6
1.85
0
NaN
0
5020
NaN
Brandon Landers
NaN
143.0
8.0
8.0
Alana Kaniewski
720.0
NaN
Drama|Horror|Thriller
...
USA
NaN
17350.0
2011-01-01
19.0
3.0
NaN
33
NaN
0
5021
Color
Jay Duplass
51.0
85.0
157.0
10.0
Katie Aselton
830.0
192467.0
Comedy|Drama|Romance
...
USA
R
15000.0
2005-01-01
224.0
6.6
NaN
297
177467.0
1
5022
Black and White
Jim Chuchu
6.0
60.0
0.0
4.0
Olwenya Maina
147.0
NaN
Drama
...
Kenya
NaN
15000.0
2014-01-01
19.0
7.4
NaN
45
NaN
0
5023
Color
Daryl Wein
22.0
88.0
38.0
211.0
Heather Burns
331.0
76382.0
Romance
...
USA
NaN
15000.0
2009-01-01
212.0
6.2
2.35
324
61382.0
1
5024
Color
Jason Trost
42.0
78.0
91.0
86.0
Jason Trost
407.0
NaN
Sci-Fi|Thriller
...
USA
Unrated
20000.0
2011-01-01
91.0
4.0
2.35
835
NaN
0
5025
Color
John Waters
73.0
108.0
0.0
105.0
Mink Stole
462.0
180483.0
Comedy|Crime|Horror
...
USA
NC-17
10000.0
1972-01-01
143.0
6.1
1.37
0
170483.0
1
5026
Color
Olivier Assayas
81.0
110.0
107.0
45.0
Béatrice Dalle
576.0
136007.0
Drama|Music|Romance
...
France
R
4500.0
2004-01-01
133.0
6.9
2.35
171
131507.0
1
5027
Color
Jafar Panahi
64.0
90.0
397.0
0.0
Nargess Mamizadeh
5.0
673780.0
Drama
...
Iran
Not Rated
10000.0
2000-01-01
0.0
7.5
1.85
697
663780.0
1
5028
Black and White
Ivan Kavanagh
12.0
83.0
18.0
0.0
Michael Parle
10.0
NaN
Horror
...
Ireland
NaN
10000.0
2007-01-01
5.0
6.7
1.33
105
NaN
0
5029
Color
Kiyoshi Kurosawa
78.0
111.0
62.0
6.0
Anna Nakagawa
89.0
94596.0
Crime|Horror|Mystery|Thriller
...
Japan
NaN
1000000.0
1997-01-01
13.0
7.4
1.85
817
-905404.0
0
5030
Color
Tadeo Garcia
NaN
84.0
5.0
12.0
Michael Cortez
21.0
NaN
Drama
...
USA
NaN
NaN
2004-01-01
20.0
6.1
NaN
22
NaN
0
5031
Color
Thomas L. Phillips
13.0
82.0
120.0
84.0
Joe Coffey
785.0
NaN
Comedy|Horror|Thriller
...
USA
NaN
200000.0
2012-01-01
98.0
5.4
16.00
424
NaN
0
5032
Color
Ash Baron-Cohen
10.0
98.0
3.0
152.0
Stanley B. Herman
789.0
NaN
Crime|Drama
...
USA
NaN
NaN
1995-01-01
194.0
6.4
NaN
20
NaN
0
5033
Color
Shane Carruth
143.0
77.0
291.0
8.0
David Sullivan
291.0
424760.0
Drama|Sci-Fi|Thriller
...
USA
PG-13
7000.0
2004-01-01
45.0
7.0
1.85
19000
417760.0
1
5034
Color
Neill Dela Llana
35.0
80.0
0.0
0.0
Edgar Tancangco
0.0
70071.0
Thriller
...
Philippines
Not Rated
7000.0
2005-01-01
0.0
6.3
NaN
74
63071.0
1
5035
Color
Robert Rodriguez
56.0
81.0
0.0
6.0
Peter Marquardt
121.0
2040920.0
Action|Crime|Drama|Romance|Thriller
...
USA
R
7000.0
1992-01-01
20.0
6.9
1.37
0
2033920.0
1
5036
Color
Anthony Vallone
NaN
84.0
2.0
2.0
John Considine
45.0
NaN
Crime|Drama
...
USA
PG-13
3250.0
2005-01-01
44.0
7.8
NaN
4
NaN
0
5037
Color
Edward Burns
14.0
95.0
0.0
133.0
Caitlin FitzGerald
296.0
4584.0
Comedy|Drama
...
USA
Not Rated
9000.0
2011-01-01
205.0
6.4
NaN
413
-4416.0
0
5038
Color
Scott Smith
1.0
87.0
2.0
318.0
Daphne Zuniga
637.0
NaN
Comedy|Drama
...
Canada
NaN
NaN
2013-01-01
470.0
7.7
NaN
84
NaN
0
5039
Color
NaN
43.0
43.0
NaN
319.0
Valorie Curry
841.0
NaN
Crime|Drama|Mystery|Thriller
...
USA
TV-14
NaN
NaT
593.0
7.5
16.00
32000
NaN
0
5040
Color
Benjamin Roberds
13.0
76.0
0.0
0.0
Maxwell Moody
0.0
NaN
Drama|Horror|Thriller
...
USA
NaN
1400.0
2013-01-01
0.0
6.3
NaN
16
NaN
0
5041
Color
Daniel Hsia
14.0
100.0
0.0
489.0
Daniel Henney
946.0
10443.0
Comedy|Drama|Romance
...
USA
PG-13
NaN
2012-01-01
719.0
6.3
2.35
660
NaN
0
5042
Color
Jon Gunn
43.0
90.0
16.0
16.0
Brian Herzlinger
86.0
85222.0
Documentary
...
USA
PG
1100.0
2004-01-01
23.0
6.6
1.85
456
84122.0
1
5043 rows × 30 columns
In [5]:
data.keyword
Out[5]:
index
plot_keywords
0
0
avatar
1
0
future
2
0
marine
3
0
native
4
0
paraplegic
5
1
goddess
6
1
marriage ceremony
7
1
marriage proposal
8
1
pirate
9
1
singapore
10
2
bomb
11
2
espionage
12
2
sequel
13
2
spy
14
2
terrorist
15
3
deception
16
3
imprisonment
17
3
lawlessness
18
3
police officer
19
3
terrorist plot
20
4
21
5
alien
22
5
american civil war
23
5
male nipple
24
5
mars
25
5
princess
26
6
sandman
27
6
spider man
28
6
symbiote
29
6
venom
...
...
...
24220
5034
security guard
24221
5034
squatter
24222
5035
assassin
24223
5035
death
24224
5035
guitar
24225
5035
gun
24226
5035
mariachi
24227
5036
jewell
24228
5036
mongol
24229
5036
nostradamus
24230
5036
stepnicka
24231
5036
vallone
24232
5037
written and directed by cast member
24233
5038
fraud
24234
5038
postal worker
24235
5038
prison
24236
5038
theft
24237
5038
trial
24238
5039
cult
24239
5039
fbi
24240
5039
hideout
24241
5039
prison escape
24242
5039
serial killer
24243
5040
24244
5041
24245
5042
actress name in title
24246
5042
crush
24247
5042
date
24248
5042
four word title
24249
5042
video camera
24250 rows × 2 columns
In [6]:
data.genre
Out[6]:
index
genres
0
0
Action
1
0
Adventure
2
0
Fantasy
3
0
Sci-Fi
4
1
Action
5
1
Adventure
6
1
Fantasy
7
2
Action
8
2
Adventure
9
2
Thriller
10
3
Action
11
3
Thriller
12
4
Documentary
13
5
Action
14
5
Adventure
15
5
Sci-Fi
16
6
Action
17
6
Adventure
18
6
Romance
19
7
Adventure
20
7
Animation
21
7
Comedy
22
7
Family
23
7
Fantasy
24
7
Musical
25
7
Romance
26
8
Action
27
8
Adventure
28
8
Sci-Fi
29
9
Adventure
...
...
...
14474
5031
Horror
14475
5031
Thriller
14476
5032
Crime
14477
5032
Drama
14478
5033
Drama
14479
5033
Sci-Fi
14480
5033
Thriller
14481
5034
Thriller
14482
5035
Action
14483
5035
Crime
14484
5035
Drama
14485
5035
Romance
14486
5035
Thriller
14487
5036
Crime
14488
5036
Drama
14489
5037
Comedy
14490
5037
Drama
14491
5038
Comedy
14492
5038
Drama
14493
5039
Crime
14494
5039
Drama
14495
5039
Mystery
14496
5039
Thriller
14497
5040
Drama
14498
5040
Horror
14499
5040
Thriller
14500
5041
Comedy
14501
5041
Drama
14502
5041
Romance
14503
5042
Documentary
14504 rows × 2 columns
In [ ]:
Content source: WaylonWalker/pyDataVizDay
Similar notebooks: