├── README.md
├── main.py
└── CPA_fig2.ipynb


/README.md:
--------------------------------------------------------------------------------
 1 | # Description of datasets:
 2 | 
 3 | All datasets were produced using the [Hooktheory API](https://www.hooktheory.com/api/trends/docs) and the [Spotify API](https://developer.spotify.com/documentation/web-api/).
 4 | 
 5 | Given a chord progression, the Hooktheory database annotates all of its child progressions with the proportion of songs containing that child node. For example, of all songs containing the progression "I, IV, vi", 55% are followed by "V" and 14% are followed by "IV". Information about interpreting the Hooktheory chord notation can be found [here](http://forum.hooktheory.com/t/vizualitation-of-all-chord-progressions-kinda/164/2).
 6 | 
 7 | Chord progressions were pulled from the Hooktheory database as follows:
 8 | * First, all one length chord progressions (single chords) which were contained in at least %5 of the Hooktheory song database were pulled.
 9 | * Next, length two chord progressions were constructed by appending to each of the length one chord progressions any chord which comprises at least 5% of all songs containing the given length one progression. So for example, 15% of songs in the Hooktheory database begin with the "I" chord. Of chords starting with "I", 6% are followed by "ii", hence the length-two chord progression "I, ii" was pulled from the database.
10 | * Continuing in this way, we extract all 3, 4, and 5-chord progressions which have at least a 5% chance of occuring given their parent progression. 
11 | 
12 | After the chord progressions were obtained, I used the Hooktheory API again to pull all songs associated with the pulled progressions. Each song item contained information about the song, artist, and section (chorus, verse, etc.) which contained the given progression.
13 | 
14 | Next, given the song/artist pairs pulled from the Hooktheory database, I used the Spotify API to query the Spotify track database to find tracks which match the song/artist pair. Of the songs for which a match was found, I then used the Spotify API again to pull detailed audio features for each track, and the genre information for the artists of the tracks. Detailed information about audio features can be found [here](https://developer.spotify.com/documentation/web-api/reference/tracks/get-audio-features/)
15 | 
16 | All of the following datasets were constructed by manipulating/cleaning the data pulled in the aforementioned manner.
17 | 
18 | ## three_, four_, and five_chord_songs.csv
19 | 
20 | As the titles suggest, these contain all three/four/five chord progressions along with any song/artist/section information pulled from the Hooktheory database. Where possible, Spotify audio feature data is given. No genre information is given.
21 | 
22 | ## three_four_five.csv
23 | 
24 | This is the concatenation of the three_, four_, five_chord_songs.csv datasets. Moreover, where possible genre information is given. The genres correspond to the artist of the song (though the song doesn't necessarily fit into each genre which describes a given artist).
25 | 
26 | ## three_four_five_pruned.csv
27 | 
28 | Several three- and four-chord progressions are contained in progressions of a longer length. For example, "I, IV, I" is contained in "I, IV, I, V" which is itself contained in "I, IV, I, V, I". This dataset removes redundant chord progressions by favoring the longer progression. Thus in the example given, if a given song/artist/section combination contains both "I, IV, I" and "I, IV, I, V", the rows pertaining to "I, IV, I" are pruned from the dataset. Similarly, if the same song/artist/section combination contains "I, IV, I, V, I", then the four-chord progression is also pruned.
29 | 
30 | ## three_four_five_has_audio_pruned.csv
31 | 
32 | This is a further refinement of three_four_five_pruned.csv which retains only those song/artist combinations for which Spotify information was successfully queried.
33 | 


--------------------------------------------------------------------------------
/main.py:
--------------------------------------------------------------------------------
  1 | #!/usr/bin/env python3
  2 | # -*- coding: utf-8 -*-
  3 | """
  4 | 2/2/18
  5 | 
  6 | Author: Jesse Hamer
  7 | 
  8 | The Data Incubator Application
  9 | 
 10 | Challenge 3: Propose a Project
 11 | 
 12 | Working Project Title: Sentiment Analysis of Chord Progressions
 13 | """
 14 | 
 15 | import requests
 16 | import time
 17 | import json
 18 | import pandas as pd
 19 | import numpy as np
 20 | import seaborn as sns
 21 | import matplotlib.pyplot as plt
 22 | 
 23 | HOOKTHEORY_USERNAME = 'your_username_here'
 24 | HOOKTHEORY_PWORD = 'your_password_here'
 25 | HOOKTHEORY_API_ENDPT = 'https://api.hooktheory.com/v1/'
 26 | 
 27 | HT_AUTH_REQUEST_URL = HOOKTHEORY_API_ENDPT + 'users/auth'
 28 | 
 29 | HT_AUTH_REQUEST_BODY = {'username': HOOKTHEORY_USERNAME,
 30 |                      'password': HOOKTHEORY_PWORD}
 31 | 
 32 | HT_AUTH_REQUEST_RESP = requests.request('POST',
 33 |                                         HT_AUTH_REQUEST_URL,
 34 |                                         json=HT_AUTH_REQUEST_BODY)
 35 | 
 36 | HT_AUTH_REQUEST_CONTENT = HT_AUTH_REQUEST_RESP.json()
 37 | 
 38 | ht_activation_key = 'Bearer ' + HT_AUTH_REQUEST_CONTENT['activkey']
 39 | 
 40 | ht_auth_header = {'Authorization': ht_activation_key}
 41 | 
 42 | client = requests.session()
 43 | client.headers = ht_auth_header
 44 | 
 45 | def get_song_request(sess, cp, wait_time=0, verbose=False):
 46 |     songs = []
 47 |     
 48 |     page = 0
 49 |     redo=True
 50 |     new_songs=[]
 51 |     
 52 |     while new_songs or redo:
 53 |         songs+=new_songs
 54 |         page+=1
 55 |         redo=False
 56 |         
 57 |         r = sess.get(HOOKTHEORY_API_ENDPT + 'trends/songs', 
 58 |                      params = {'cp':cp, 'page':str(page)})
 59 |         new_songs = r.json()
 60 |         
 61 |         remaining = int(r.headers['X-Rate-Limit-Remaining'])
 62 |         wait_time = int(r.headers['X-Rate-Limit-Reset'])
 63 |         
 64 |         if verbose:
 65 |             print('Retrieved page {}; contains {} new results'.format(page,
 66 |                   len(new_songs)))
 67 |             
 68 |         if remaining==0:
 69 |             print('Too many requests. Waiting {} seconds...'.format(wait_time))
 70 |             time.sleep(wait_time)
 71 |         if False in [type(s)==dict for s in new_songs]:
 72 |             page-=1
 73 |             new_songs=[]
 74 |             redo=True
 75 |             
 76 |     return songs
 77 | 
 78 | 
 79 | def get_chord_progressions(sess, initial_progressions, tol = 0, verbose=False):
 80 |     chord_progs = []
 81 |     initial_progs = initial_progressions.copy()
 82 |     
 83 |     while initial_progs:
 84 |         prog = initial_progs.pop(0)
 85 |         cp= prog['child_path']
 86 |         
 87 |         r = sess.get(HOOKTHEORY_API_ENDPT + 'trends/nodes', 
 88 |                      params = {'cp':cp})
 89 |         new_prog = r.json()
 90 |         
 91 |         remaining = int(r.headers['X-Rate-Limit-Remaining'])
 92 |         wait_time = int(r.headers['X-Rate-Limit-Reset'])
 93 |             
 94 |         if remaining==0:
 95 |             print('Too many requests. Waiting {} seconds...'.format(wait_time))
 96 |             time.sleep(wait_time)
 97 |         if False in [type(s)==dict for s in new_prog]:
 98 |             initial_progs.insert(0, prog)
 99 |         else:
100 |             new_prog = [p for p in new_prog if p['probability']>tol]
101 |             chord_progs+=new_prog
102 |             if verbose:
103 |                 print('Progression {} processed.'.format(cp))
104 |             
105 |             
106 |     return chord_progs
107 | 
108 | one_chord = get_chord_progressions(client, [{'child_path':''}], tol=0.05, verbose=True)
109 | 
110 | two_chord = get_chord_progressions(client, one_chord, tol=0.05, verbose=True)
111 | 
112 | three_chord = get_chord_progressions(client, two_chord, tol=0.05, verbose=True)
113 | 
114 | four_chord = get_chord_progressions(client, three_chord, tol=0.05, verbose=True)
115 | 
116 | five_chord = get_chord_progressions(client, four_chord, tol=0.05, verbose=True)
117 | 
118 | def get_cp_song_data(sess, chord_progs, verbose=0):
119 |     data = pd.DataFrame([], columns=['cp', 'artist', 'song', 'section'])
120 |     total_cp = len(chord_progs)
121 |     cp_counter = 0
122 |     
123 |     if verbose==0:
124 |         songs_verbose=False
125 |         cp_verbose=False
126 |     if verbose==1:
127 |         songs_verbose=False
128 |         cp_verbose=True
129 |     if verbose==2:
130 |         songs_verbose=True
131 |         cp_verbose=True
132 |     
133 |     for prog in chord_progs:
134 |         cp_counter+=1
135 |         cp = prog['child_path']
136 |         if cp_verbose:
137 |             print('###### FETCHING SONGS FOR {}; {}/{} ##########\n'.format(cp,
138 |                   cp_counter, total_cp))
139 |         songs = get_song_request(sess, cp, verbose=songs_verbose)
140 |         for song in songs:
141 |             song['cp'] = cp
142 |             song.pop('url')
143 |         data = data.append(songs)
144 |         if cp_verbose:
145 |             print('###### DONE WITH {}; {} SONGS ADDED; DATA SHAPE: {} #########\n'.format(cp,
146 |                   len(songs), data.shape))
147 |     return data
148 | 
149 | # Get data for four-chord progressions first, to make sure everything works.
150 |         
151 | cp_song_data_four = get_cp_song_data(client, four_chord, verbose=1)
152 | 
153 | cp_song_data_four.song = cp_song_data_four.song.apply(lambda x: x.lower())
154 | cp_song_data_four.artist = cp_song_data_four.artist.apply(lambda x: x.lower())
155 | cp_song_data_four.section= cp_song_data_four.section.apply(lambda x: x.lower())
156 | 
157 | print(cp_song_data_four.describe())
158 | 
159 | # 2220 unique artists, 3746 unique songs; 22 unique sections
160 | 
161 | print(cp_song_data_four.groupby(['artist', 'song']).size().shape[0])
162 | 
163 | # 3874 unique artist/song combinations
164 | 
165 | cp4_artist_song = cp_song_data_four[['artist', 'song']]
166 | cp4_artist_song.drop_duplicates(inplace=True)
167 | cp4_artist_song.sort_values(by=['artist', 'song'], inplace=True)
168 | cp4_artist_song.reset_index(inplace=True, drop=True)
169 | 
170 | 
171 | 
172 | ##############################################
173 | 
174 | # Now try to retrieve spotify information for each artist/song    
175 | 
176 | from spotipy import Spotify
177 | from spotipy.oauth2 import SpotifyClientCredentials
178 | 
179 | SPOTIFY_CLIENT_ID = 'your_spotify_client_id_here'
180 | 
181 | SPOTIFY_CLIENT_SECRET = 'your_spotify_client_secret_here'
182 | 
183 | token = SpotifyClientCredentials(client_id=SPOTIFY_CLIENT_ID,
184 |                                  client_secret=SPOTIFY_CLIENT_SECRET)
185 | 
186 | cache_token = token.get_access_token()
187 | 
188 | spotify = Spotify(cache_token)
189 | 
190 | def get_track_ids(client, data, num_tracks=None, turn_update=None):
191 |     
192 |     for i, row in data.iterrows():
193 |         artist=row['artist']
194 |         song=row['song']
195 |         q = 'artist:'+artist + ' ' + 'track:'+song
196 |         result = client.search(q)
197 |         items = result['tracks']['items']
198 |         if not items:
199 |             continue
200 |         popularities = [item['popularity'] for item in items]
201 |         most_popular = items[popularities.index(max(popularities))]
202 |         data.loc[(data.artist==artist) & (data.song==song),'spotify_ID']=most_popular['id']
203 |         if i == num_tracks:
204 |             break
205 |         if turn_update and (i+1)%turn_update==0:
206 |             print('Finished track {}'.format(i+1))
207 |             
208 | get_track_ids(spotify, cp4_artist_song)
209 |     
210 | def get_audio_features(client, data, verbose=False):
211 |     
212 |     ids = data.spotify_ID.dropna()
213 |     
214 |     audio_feature_data = pd.DataFrame([],columns=['danceability', 'energy',
215 |                                       'key', 'loudness', 'mode', 'speechiness',
216 |                                       'acousticness', 'instrumentalness',
217 |                                       'liveness', 'valence', 'tempo','id',
218 |                                       'duration_ms', 'time_signature'])
219 |     for i in range(0, len(ids), 50):
220 |         new_features = spotify.audio_features(ids[i:i+50])
221 |         for track in new_features:
222 |             track.pop('type')
223 |             track.pop('uri')
224 |             track.pop('track_href')
225 |             track.pop('analysis_url')
226 |         audio_feature_data = audio_feature_data.append(new_features)
227 |         if verbose:
228 |             print('Done with tracks {} through {}'.format(i+1, i+50))
229 |     
230 |     return audio_feature_data
231 |     
232 | cp4_audio_features = get_audio_features(spotify, cp4_artist_song,verbose=True)    
233 | 
234 | cp4_audio_features.rename(columns={'id':'spotify_ID'}, inplace=True)        
235 | 
236 | cp4_artist_song = cp4_artist_song.merge(cp4_audio_features, how='left', 
237 |                                         on='spotify_ID')
238 | 
239 | cp_song_data_four = cp_song_data_four.merge(cp4_artist_song, how='left',
240 |                                             on=['artist','song'])
241 | 
242 | cp_song_data_four.to_csv('four_chord_songs.csv', index=False)
243 | 
244 | del cp_song_data_four
245 | del cp4_artist_song
246 | del cp4_audio_features
247 | 
248 | 
249 | ##################################
250 | # Now repeat for three and five chord progressions.
251 | ##################################
252 | 
253 | # THREE CHORD PROGRESSIONS:
254 | 
255 | HT_AUTH_REQUEST_RESP = requests.request('POST',
256 |                                         HT_AUTH_REQUEST_URL,
257 |                                         json=HT_AUTH_REQUEST_BODY)
258 | 
259 | HT_AUTH_REQUEST_CONTENT = HT_AUTH_REQUEST_RESP.json()
260 | 
261 | ht_activation_key = 'Bearer ' + HT_AUTH_REQUEST_CONTENT['activkey']
262 | 
263 | ht_auth_header = {'Authorization': ht_activation_key}
264 | 
265 | client = requests.session()
266 | client.headers = ht_auth_header
267 | 
268 | cp_song_data_three = get_cp_song_data(client, three_chord, verbose=1)
269 | 
270 | cp_song_data_three.song = cp_song_data_three.song.apply(lambda x: x.lower())
271 | cp_song_data_three.artist = cp_song_data_three.artist.apply(lambda x: x.lower())
272 | cp_song_data_three.section= cp_song_data_three.section.apply(lambda x: x.lower())
273 | 
274 | 
275 | cp3_artist_song = cp_song_data_three[['artist', 'song']]
276 | cp3_artist_song.drop_duplicates(inplace=True)
277 | cp3_artist_song.sort_values(by=['artist', 'song'], inplace=True)
278 | cp3_artist_song.reset_index(inplace=True, drop=True)
279 | 
280 | token = SpotifyClientCredentials(client_id=SPOTIFY_CLIENT_ID,
281 |                                  client_secret=SPOTIFY_CLIENT_SECRET)
282 | 
283 | cache_token = token.get_access_token()
284 | 
285 | spotify = Spotify(cache_token)
286 | 
287 | get_track_ids(spotify, cp3_artist_song)
288 | 
289 | cp3_audio_features = get_audio_features(spotify, cp3_artist_song,verbose=True)    
290 | 
291 | cp3_audio_features.rename(columns={'id':'spotify_ID'}, inplace=True)        
292 | 
293 | cp3_artist_song = cp3_artist_song.merge(cp3_audio_features, how='left', 
294 |                                         on='spotify_ID')
295 | 
296 | cp_song_data_three = cp_song_data_three.merge(cp3_artist_song, how='left',
297 |                                             on=['artist','song'])
298 | 
299 | cp_song_data_three.to_csv('three_chord_songs.csv', index=False)
300 | 
301 | del cp_song_data_three
302 | del cp3_artist_song
303 | del cp3_audio_features
304 | 
305 | ######################################
306 | 
307 | # FIVE CHORD PROGRESSIONS
308 | 
309 | HT_AUTH_REQUEST_RESP = requests.request('POST',
310 |                                         HT_AUTH_REQUEST_URL,
311 |                                         json=HT_AUTH_REQUEST_BODY)
312 | 
313 | HT_AUTH_REQUEST_CONTENT = HT_AUTH_REQUEST_RESP.json()
314 | 
315 | ht_activation_key = 'Bearer ' + HT_AUTH_REQUEST_CONTENT['activkey']
316 | 
317 | ht_auth_header = {'Authorization': ht_activation_key}
318 | 
319 | client = requests.session()
320 | client.headers = ht_auth_header
321 | 
322 | cp_song_data_five = get_cp_song_data(client, five_chord, verbose=1)
323 | 
324 | cp_song_data_five.song = cp_song_data_five.song.apply(lambda x: x.lower())
325 | cp_song_data_five.artist = cp_song_data_five.artist.apply(lambda x: x.lower())
326 | cp_song_data_five.section= cp_song_data_five.section.apply(lambda x: x.lower())
327 | 
328 | 
329 | cp5_artist_song = cp_song_data_five[['artist', 'song']]
330 | cp5_artist_song.drop_duplicates(inplace=True)
331 | cp5_artist_song.sort_values(by=['artist', 'song'], inplace=True)
332 | cp5_artist_song.reset_index(inplace=True, drop=True)
333 | 
334 | token = SpotifyClientCredentials(client_id=SPOTIFY_CLIENT_ID,
335 |                                  client_secret=SPOTIFY_CLIENT_SECRET)
336 | 
337 | cache_token = token.get_access_token()
338 | 
339 | spotify = Spotify(cache_token)
340 | 
341 | get_track_ids(spotify, cp5_artist_song)
342 | 
343 | cp5_audio_features = get_audio_features(spotify, cp5_artist_song,verbose=True)    
344 | 
345 | cp5_audio_features.rename(columns={'id':'spotify_ID'}, inplace=True)        
346 | 
347 | cp5_artist_song = cp5_artist_song.merge(cp5_audio_features, how='left', 
348 |                                         on='spotify_ID')
349 | 
350 | cp_song_data_five = cp_song_data_five.merge(cp5_artist_song, how='left',
351 |                                             on=['artist','song'])
352 | 
353 | cp_song_data_five.to_csv('five_chord_songs.csv', index=False)
354 | 
355 | del cp_song_data_five
356 | del cp5_artist_song
357 | del cp5_audio_features
358 | 
359 | ###########################################
360 | 
361 | # Get genre information for all tracks
362 | 
363 | three = pd.read_csv('three_chord_songs.csv')
364 | four = pd.read_csv('four_chord_songs.csv')
365 | five = pd.read_csv('five_chord_songs.csv')
366 | 
367 | three['cp_length'] = 3
368 | four['cp_length'] = 4
369 | five['cp_length'] = 5
370 | 
371 | three_four_five = three.append(four).append(five).reset_index(drop=True)
372 | 
373 | def get_track_genres(client, track_ids, verbose=False):
374 |     data = pd.DataFrame([], columns=['spotify_ID', 'genres'])
375 |     for i in range(0, len(track_ids), 20):
376 |         tids = track_ids[i:i+20]
377 |         tracks = client.tracks(tids)['tracks']
378 |         artist_ids = [track['artists'][0]['id'] for track in tracks]
379 |         artists = client.artists(artist_ids)['artists']
380 |         genres = [artist['genres'] for artist in artists]
381 |         new_data = [{'spotify_ID':tid, 'genres':genre} for tid,genre in zip(tids, genres)]
382 |         data = data.append(new_data).dropna()
383 |         if verbose:
384 |             print('Finished fetching genres for tids {} through {}.'.format(i+1, i+20))
385 |             print('New data shape: {}'.format(data.shape))
386 |             
387 |     return data
388 |         
389 | token = SpotifyClientCredentials(client_id=SPOTIFY_CLIENT_ID,
390 |                                  client_secret=SPOTIFY_CLIENT_SECRET)
391 | 
392 | cache_token = token.get_access_token()
393 | 
394 | spotify = Spotify(cache_token)
395 | 
396 | track_genres = get_track_genres(spotify, three_four_five.spotify_ID.dropna().unique(),
397 |                                 verbose=True)
398 | 
399 | three_four_five = three_four_five.merge(track_genres, how='left', on='spotify_ID')
400 | 
401 | three_four_five.to_csv('three_four_five.csv', index=False)
402 | 
403 | # If one chord progression is contained within another, we favor the longer.
404 | def remove_redundant_cp(data, l1, l2):
405 |     cp_l1 = data.loc[data.cp_length==l1,['cp', 'artist', 'song', 'section']]
406 |     cp_l2 = data.loc[data.cp_length==l2, ['cp', 'artist', 'song', 'section']]
407 |     
408 |     for i, row in cp_l1.iterrows():
409 |         cp = row['cp']
410 |         artist=row['artist']
411 |         song=row['song']
412 |         section=row['section']
413 |         if ((cp_l2.cp.apply(lambda x: cp in x))&((artist==cp_l2.artist)&\
414 |              ((song==cp_l2.song)&(section==cp_l2.section)))).any():
415 |             data.drop(i, inplace=True)
416 | 
417 | three_four_five_pruned = three_four_five.copy()
418 | 
419 | remove_redundant_cp(three_four_five_pruned, 3, 4)
420 | remove_redundant_cp(three_four_five_pruned, 3, 5)    
421 | remove_redundant_cp(three_four_five_pruned, 4, 5)
422 | 
423 | # Save the new 'pruned' datasets
424 | 
425 | three_four_five_pruned.to_csv('three_four_five_pruned.csv', index=False)
426 | 
427 | # Make dataset of cp/artist/song/section combinations which have spotify info
428 | 
429 | has_audio_data_pruned = three_four_five_pruned.dropna(subset=['spotify_ID'])
430 | 
431 | has_audio_data_pruned.to_csv('three_four_five_has_audio_pruned.csv', index=False)
432 | 
433 | has_audio_data_pruned.reset_index(drop=True, inplace=True)
434 | 
435 | ####################################
436 | 
437 | # NOW FOR EDA
438 | 
439 | print(three_four_five.shape)
440 | 
441 | print(has_audio_data_pruned.shape)
442 | 
443 | # 12511 non-redundant records with audio data
444 | 
445 | print(has_audio_data_pruned.cp_length.value_counts())
446 | 
447 | # 9884 5-chord progressions, 1479 3-chord progressions, 1148 4-chord progressions
448 | 
449 | print(has_audio_data_pruned.cp.describe()) 
450 | 
451 | # 1018 unique chord progressions, 
452 | 
453 | print(has_audio_data_pruned.groupby('cp_length').apply(lambda x: x.cp.describe()))
454 | 
455 | # 68 unique 3-chord progressions, 197 unique 4-chord progressions, 753 unique
456 | # 5-chord progressions
457 | 
458 | # Note: to reformat genres as lists after loading csv, run the following:
459 | 
460 | # has_audio_data_pruned.genres = has_audio_data_pruned.genres.apply(
461 | #    lambda x: [g.strip("' ") for g in x.strip('[]').split(',')] if x!='[]' else np.nan)
462 | 
463 | # Get all genres
464 | all_genres = {}
465 | 
466 | for genres in has_audio_data_pruned.genres.dropna():
467 |     for g in genres:
468 |         if g in all_genres.keys():
469 |             all_genres[g]+=1
470 |         else:
471 |             all_genres[g]=1
472 | 
473 | all_genres = pd.Series(all_genres)
474 | 
475 | print('The 20 most popular genres are: \n{}'.format(all_genres.sort_values(ascending=False).head(20)))
476 | """
477 | pop                 3016
478 | rock                2499
479 | dance pop           2483
480 | pop rock            1589
481 | modern rock         1511
482 | post-teen pop       1393
483 | edm                 1074
484 | permanent wave       947
485 | pop punk             940
486 | album rock           930
487 | folk-pop             915
488 | mellow gold          810
489 | indie rock           755
490 | soft rock            739
491 | classic rock         729
492 | indie pop            721
493 | alternative rock     696
494 | post-grunge          685
495 | hard rock            633
496 | neo mellow           631
497 | """
498 | 
499 | # Only use cps with at least 5 observations:
500 | 
501 | cp_group_sizes = has_audio_data_pruned.groupby('cp').size()
502 | cp_group_sizes.name='n'
503 | cp_group_sizes = cp_group_sizes.reset_index()
504 | 
505 | 
506 | has_audio_data_pruned = has_audio_data_pruned.merge(cp_group_sizes, on='cp')
507 | 
508 | has_5_obs = has_audio_data_pruned[has_audio_data_pruned.n>=5]
509 | 
510 | print('Still have {} unique chord_progressions.'.format(has_5_obs.cp.unique().shape))
511 | 
512 | print(has_5_obs.groupby('cp_length').apply(lambda x: x.cp.describe()))
513 | 
514 | """
515 | cp         count  unique        top  freq
516 | cp_length                                
517 | 3           1473      66      4,5,1   104
518 | 4            932      93    6,4,1,5    29
519 | 5           9066     342  4,1,5,6,4   312
520 | """
521 | 
522 | all_genres = {}
523 | 
524 | for genres in has_5_obs.genres:
525 |     for g in genres:
526 |         if g in all_genres.keys():
527 |             all_genres[g]+=1
528 |         else:
529 |             all_genres[g]=1
530 | 
531 | all_genres = pd.Series(all_genres)
532 | print('The 20 most popular genres are: \n{}'.format(all_genres.sort_values(ascending=False).head(20)))
533 | """
534 | The 20 most popular genres are: 
535 | pop                 2820
536 | dance pop           2306
537 | rock                2280
538 | pop rock            1495
539 | modern rock         1366
540 | post-teen pop       1308
541 | edm                  972
542 | pop punk             893
543 | permanent wave       885
544 | album rock           852
545 | folk-pop             846
546 | mellow gold          758
547 | soft rock            688
548 | classic rock         680
549 | indie rock           670
550 | indie pop            643
551 | post-grunge          642
552 | alternative rock     621
553 | neo mellow           596
554 | hard rock            567
555 | """
556 | 
557 | numeric_audio_features = ['danceability', 'energy', 'loudness', 
558 |                            'acousticness', 'valence', 'tempo']
559 | 
560 | def all_cp_plot(data, features):
561 |     
562 |     num_plots = len(features)
563 |     cols = int(np.ceil(num_plots/3))
564 |     fig = plt.figure(figsize=(6*cols,9))
565 |     fig.subplots_adjust(hspace=.5, wspace=.3)
566 |     
567 |     for i, feature in enumerate(features):
568 |         ax = fig.add_subplot(3, cols, i+1)
569 |         cp_feature_groups = data.groupby('cp')[feature].agg([np.mean, np.std])
570 |         feature_sorted = cp_feature_groups.sort_values(by='mean')
571 |         feature_sorted['mean'].plot(ax=ax)
572 |         plt.fill_between(feature_sorted.index, 
573 |                          feature_sorted['mean']-feature_sorted['std'],
574 |                          feature_sorted['mean'] + feature_sorted['std'], 
575 |                          color='orange', alpha=0.2)
576 |         plt.xticks([],[])
577 |         plt.ylabel(feature)
578 |         ax.set_title('Mean {}'.format(feature))
579 |     plt.show()
580 | 
581 |     
582 | def cp_plot(cp, data, numeric_features=[], compare=False):
583 |     cp_data = data[data.cp==cp]
584 |     num_plots = len(numeric_features)
585 |     cols = int(np.ceil(num_plots/3))
586 |     fig = plt.figure(figsize=(6*cols,9))
587 |     fig.subplots_adjust(hspace=.5, wspace=.3)
588 |     for i, feature in enumerate(numeric_features):
589 |         ax = fig.add_subplot(3, cols, i+1)
590 |         sns.distplot(cp_data[feature], hist=False, ax=ax, label=cp)
591 |         if compare:
592 |             sns.distplot(data[feature], color='orange', hist=False, ax=ax,
593 |                          label='All CPs')
594 |         ax.set_title('Distribution of {} for {}'.format(feature, cp))
595 |     plt.show()
596 |         
597 | 
598 | 
599 | 
600 | 
601 | 
602 | 
603 | 
604 | 
605 | 
606 | 
607 | 
608 | 
609 | 
610 | 


--------------------------------------------------------------------------------
/CPA_fig2.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "markdown",
  5 |    "metadata": {},
  6 |    "source": [
  7 |     "# Audio Features for Single Chord Progression\n",
  8 |     "\n",
  9 |     "This figure compares one of the highest-valence (happiest) chord progressions to all other chord progressions."
 10 |    ]
 11 |   },
 12 |   {
 13 |    "cell_type": "code",
 14 |    "execution_count": 1,
 15 |    "metadata": {},
 16 |    "outputs": [],
 17 |    "source": [
 18 |     "import pandas as pd\n",
 19 |     "import seaborn as sns\n",
 20 |     "import matplotlib.pyplot as plt\n",
 21 |     "import numpy as np\n",
 22 |     "\n",
 23 |     "%matplotlib inline\n",
 24 |     "\n",
 25 |     "# load data\n",
 26 |     "\n",
 27 |     "has_audio_data_pruned = pd.read_csv('three_four_five_has_audio_pruned.csv')\n",
 28 |     "\n",
 29 |     "cp_group_sizes = has_audio_data_pruned.groupby('cp').size()\n",
 30 |     "cp_group_sizes.name='n'\n",
 31 |     "cp_group_sizes = cp_group_sizes.reset_index()\n",
 32 |     "\n",
 33 |     "\n",
 34 |     "has_audio_data_pruned = has_audio_data_pruned.merge(cp_group_sizes, on='cp')\n",
 35 |     "\n",
 36 |     "# We only look at chord progressions associated to at least five artist/song/section combinations\n",
 37 |     "has_5_obs = has_audio_data_pruned[has_audio_data_pruned.n>=5]"
 38 |    ]
 39 |   },
 40 |   {
 41 |    "cell_type": "code",
 42 |    "execution_count": 3,
 43 |    "metadata": {},
 44 |    "outputs": [
 45 |     {
 46 |      "name": "stderr",
 47 |      "output_type": "stream",
 48 |      "text": [
 49 |       "/Applications/anaconda3/lib/python3.6/site-packages/scipy/stats/stats.py:1713: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
 50 |       "  return np.add.reduce(sorted[indexer] * weights, axis=axis) / sumval\n"
 51 |      ]
 52 |     },
 53 |     {
 54 |      "data": {
 55 |       "image/png": "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\n",
 56 |       "text/plain": [
 57 |        "<Figure size 864x648 with 6 Axes>"
 58 |       ]
 59 |      },
 60 |      "metadata": {},
 61 |      "output_type": "display_data"
 62 |     }
 63 |    ],
 64 |    "source": [
 65 |     "numeric_audio_features = ['danceability', 'energy', 'loudness', \n",
 66 |     "                           'acousticness', 'valence', 'tempo']\n",
 67 |     "\n",
 68 |     "\n",
 69 |     "def cp_plot(cp, data, numeric_features=[], compare=False):\n",
 70 |     "    cp_data = data[data.cp==cp]\n",
 71 |     "    num_plots = len(numeric_features)\n",
 72 |     "    cols = int(np.ceil(num_plots/3))\n",
 73 |     "    fig = plt.figure(figsize=(6*cols,9))\n",
 74 |     "    fig.subplots_adjust(hspace=.5, wspace=.3)\n",
 75 |     "    for i, feature in enumerate(numeric_features):\n",
 76 |     "        ax = fig.add_subplot(3, cols, i+1)\n",
 77 |     "        sns.distplot(cp_data[feature], hist=False, ax=ax, label=cp)\n",
 78 |     "        if compare:\n",
 79 |     "            sns.distplot(data[feature], color='orange', hist=False, ax=ax,\n",
 80 |     "                         label='All CPs')\n",
 81 |     "        ax.set_title('Distribution of {} for {}'.format(feature, cp))\n",
 82 |     "    plt.show()\n",
 83 |     "    \n",
 84 |     "cp_plot('1,4,1,5,1',has_5_obs, numeric_features=numeric_audio_features, compare=True)"
 85 |    ]
 86 |   }
 87 |  ],
 88 |  "metadata": {
 89 |   "kernelspec": {
 90 |    "display_name": "Python 3",
 91 |    "language": "python",
 92 |    "name": "python3"
 93 |   },
 94 |   "language_info": {
 95 |    "codemirror_mode": {
 96 |     "name": "ipython",
 97 |     "version": 3
 98 |    },
 99 |    "file_extension": ".py",
100 |    "mimetype": "text/x-python",
101 |    "name": "python",
102 |    "nbconvert_exporter": "python",
103 |    "pygments_lexer": "ipython3",
104 |    "version": "3.6.8"
105 |   }
106 |  },
107 |  "nbformat": 4,
108 |  "nbformat_minor": 2
109 | }
110 | 


--------------------------------------------------------------------------------