Total Tracks: 450225
New Songs today: 170
Scheduled Songs: 1
API Requests: 9398
The histogram shows the root count of songs in dependence of their score.
About half of the songs are having a score of 0 while only 135 songs have a score larger than 20.
The line shows the growth of the database in songs per day.
The first part is pretty much linear until I implemented another method to find "good" songs and it started to accelerated quickly.
At some point I ran into a bug that caused the database to die for a weekend and me shutting down the crawler for a bit resulting in a small plateau. Afterwards I slowed the crawler down to prevent further issues.
The blots show almost no correlation whatsoever.
There is a slight positive correlation between score and popularity and score which i would have expected to be larger.
To my surprise I could not find a correlation between energy and popularity.
Looking at the values of danceability I can see a peak around 80/81 with two to five times as many songs as the surrounding values, hinting there could be a bug in spotifys analyzer.
Danceability vs. Popularity
Pearsons R = -0.087
Energy vs. Danceability
Pearsons R = -0.274
Energy vs. Popularity
Pearsons R = -0.016
Popularity vs. Score
Pearsons R = 0.282
Danceability vs. Score
Pearsons R = 0.054
Energy vs. Score
Pearsons R = 0.026