Spotify Popularity Model

This notebook details the building and tuning of a machine learning model that predicts a song’s popularity on Spotify.

In the music industry, an important metric that has surfaced in the last few years is an artist's Spotify numbers. Many entities in the industry, from venues to record labels, will check an artist's Spotify numbers before choosing to work with or book said artist and, for better or worse, will base their decision in part on the artist's performance on the platform.

This increase in Spotify metric importance has opened opportunities for optimization in the pop and Nashville country music workflow. Typically, a producer or songwriter will rent studio time and hire studio musicians to produce singles that can then be pitched to artists. Artists buy these songs and rerecord them with their own studio teams to be released as singles or as part of a record.

If producers had a model that they could use to evaluate their music while in the production and could deliver a model's predictions while pitching music, they would have a new edge in the industry.

The project can be viewed in detail here.