Monotonous Music?
I used
entropy over song lyrics as a proxy for repetitiveness in Billboard hits (1950-2015).
Lower entropy means more repeated words; higher entropy means more lyrical variety.
Dataset source: kevinschaich/billboard.
Examples from the results include low-entropy tracks like Harlem Shake
and high-entropy tracks like 6 Foot 7 Foot.
- Tokenize words in each song's lyrics.
- Compute word probabilities for each track.
- Compute entropy and rank songs/artists/genres by that score.
Results
Most Repetitive Billboard Songs
Most Non-Repetitive Billboard Songs
Artists With Most Repetitive Average Output
Artists With Most Non-Repetitive Average Output
Genre Ranking (Most to Least Diverse Lyrics)
| # |
Genre |
| 1 | Rap and Hip Hop |
| 2 | Folk |
| 3 | Blues |
| 4 | Rock |
| 5 | Country |
| 6 | Pop |
| 7 | Jazz |
Higher rank here means higher average entropy (less repetition).
How Repetitiveness Changed Over Time
