Analyzing Audio Features Across Genres to Understand Song Popularity: A Data-Driven Approach Using Spotify API

Abstract

This project investigates an overview of which main audio features define the popularity of any song genre. The dataset was gathered through the Spotify API and was tested using statistical methods and correlations. In this paper, 1,000 tracks of four genres are used: pop, rock, hip-hop, and country. Statistical methods of t-test and correlation matrix have been applied to key features: tempo, energy, danceability, valence, and acousticness. These features were then used to identify significant patterns and trends related to song popularity. According to the findings, tempo, energy, and valence result to be the most correlated features with song popularity, although these relations are quite different across genres.