casrain.blogg.se

One note is that each feature has been scaled to have a minimum of zero and a maximum of one. In the plot below, we graph the average of each feature taken of the 100 songs in each playlist for me and for my girlfriend as well. Notes: length is taken from duration_ms on page (1) and release_year is the year element of release_date on page (1): You can read more about them at these locations. These include the features: acousticness, energy, danceability, valence, liveness, speechiness, instrumentalness, loudness, length, tempo, popularity, mode, key, time_signature, and release_year for each song. are dropped from the end of a word before plottingĪnother interesting visualization to consider contains musical features that are offered by Spotify’s API. have been removed since they are more common and do not provide as much meaning

One of the largest words, feat, shows that a good amount of songs are featuring other artists. In addition, the word time in the bottom right shows that many songs are about feelings over time. For example, the word friend in the top left corner can show that some songs are about friendship. This offers an idea of the meaning of the songs. In this image, the size of the word represents how often it shows up. The total numbers accumulate as time goes on. The first graph below is a bar plot that shows my top 10 artists based on the number of their songs I had on my playlists in 2016-2021. In addition to looking at the songs per artist in a table, we can also look at it in a graph. And to see how many artists total were on each playlist: len_artists = len(df_2016].groupby('artist').count()) print(f'You had artists in the 100 songs in 2021.') You had 36 artists in the 100 songs in 2021.