Question:
What early signals predict sustained success in the music industry?

Objective:
Many musicians produce hit songs, but not all are able to do so more than once. This project builds a machine learning classifier to distinguish hitmakers (artists with multiple top 20 Billboard Hot 100 hits) from one-hit wonders, using only information available at the moment of a musician’s first top 20 hit song.

Conclusions:
Our model reveals that prior charting experience, collaboration network position, chart longevity, genre breadth, and dominant genre affiliations are the strongest predictors of sustained success.

Data sources:
- MusicBrainz (artist metadata, genre tags, collaboration graph)
- Billboard Hot 100 & 200 chart data
- Spotify (artist and song metadata)
- Google Trends (relative search volume at time of first hit song)