Abstract

This study approached the Hit Song Science problem with the aim of predictingwhich songs in the Afrobeats genre will become popular among Spotify listeners.A dataset of 2063 songs was generated through the Spotify Web API, with theprovided audio features. Random Forest and Gradient Boosting algorithms provedto be successful with approximately F1 scores of 86%.

 

Quick Read (beta)