zeppelin_audio_features.R

library(httr)

country = "GB"
albumType = "album"
artist="Led Zeppelin"

#get artist id
url <- paste0("https://api.spotify.com/v1/search?q=", sub(" ", "%20", artist), "&type=artist")
search <- content(GET(url))
search$artists$items[[1]]
artist_id = search$artists$items[[1]]$id

#user artis id to get list of albums ids
url <- paste0("https://api.spotify.com/v1/artists/", artist_id, "/albums?album_type=", albumType, "&limit=50&country=", country)
list <- content(GET(url))

albums <- data.frame(artist = NULL, artist_id = NULL, album = NULL,  album_id = NULL)

for (j in c(1:length(list$items))) {
    temp <- data.frame(artist = artist, 
                       artist_id = artist_id, 
                       album = list$items[[j]]$name, 
                       album_id = list$items[[j]]$id)
    albums <- rbind(albums, temp)
}

#pick out original studio albums
studio_albums<-albums[c(34, 31, 28, 25, 21, 15, 14, 10, 7),]

#for each album get track ids
albumtracks <- data.frame(artist = NULL, artist_id = NULL, 
                          album = NULL, album_id = NULL, track = NULL, track_id = NULL, 
                          track_number = NULL, track_length = NULL, preview_url = NULL)

for (i in c(1:nrow(studio_albums))) {
  url <- paste0("https://api.spotify.com/v1/albums/", studio_albums[i,4], "/tracks?limit=50")
  list <- content(GET(url))
  for (j in c(1:length(list$items))) {
    temp <- data.frame(artist = studio_albums[i,1], artist_id = studio_albums[i, 2], 
                       album = studio_albums[i, 3], album_id = studio_albums[i, 4],
                       track = list$items[[j]]$name, track_id = list$items[[j]]$id, 
                       track_number = list$items[[j]]$track_number, 
                       track_length = format(.POSIXct(list$items[[j]]$duration_ms/1000, tz = "GMT"), "%M:%S"), 
                       preview_url = ifelse(is.null(list$items[[j]]$preview_url), 
                       "NO PREVIEW", list$items[[j]]$preview_url)
                       )
    albumtracks <- rbind(albumtracks, temp)
  }
}

albumtracks$track_id[1]



#for each track id get audio features. need to be a registered spotify dev with id and access key

clientID = 'xxxxxxxxxxxxxxxx'
secret = 'xxxxxxxxxxxxxxxxxxxxxx'

#get token
response = POST(
  'https://accounts.spotify.com/api/token',
  accept_json(),
  authenticate(clientID, secret),
  body = list(grant_type = 'client_credentials'),
  encode = 'form',
  verbose()
)

mytoken = content(response)$access_token
HeaderValue = paste0('Bearer ', mytoken)



#get features
audio_features<-data.frame(
  danceability=NULL,  energy=NULL,  key=NULL,  loudness=NULL,  mode=NULL,  speechiness=NULL,
  acousticness=NULL,  instrumentalness=NULL,  liveness=NULL,  valence=NULL,  tempo=NULL,
  duration_ms=NULL,  time_signature=NULL
)

for(i in albumtracks$track_id){
  url <- paste0("https://api.spotify.com/v1/audio-features/", i)
  list = content(GET(url = url, add_headers(Authorization = HeaderValue)))
  temp<-data.frame(
    danceability=list$danceability[1],
    energy=list$energy[1],
    key=list$key[1],
    loudness=list$loudness[1],
    mode=list$mode[1],
    speechiness=list$speechiness[1],
    acousticness=list$acousticness[1],
    instrumentalness=list$instrumentalness[1],
    liveness=list$liveness[1],
    valence=list$valence[1],
    tempo=list$tempo[1],
    duration_ms=list$duration_ms[1],
    time_signature=list$time_signature[1]
  )
  audio_features<-rbind(audio_features, temp)
}

#combine and save data

albumtrackfeatures<-cbind(albumtracks, audio_features)

zep_output<-list(albumtrackfeatures=albumtrackfeatures, audio_features=audio_features)

save(zep_output, file="zep_tracks.RData")

write.table(albumtrackfeatures, file="zep_tracks.csv", row.names=FALSE, col.names = FALSE, sep=",")