Spotify Popularity Analysis — Academic Project (2025)

Spotify Song Popularity Analysis Academic project · University of Toronto · 2025

This academic project examined factors associated with Spotify song popularity using a dataset of 953 tracks. I built multiple linear regression models in R and carried out data cleaning, variable transformations, residual diagnostics, and model validation.

The analysis compared production-related characteristics with emotional features to assess how each related to streaming popularity. Within the dataset, production-related characteristics were stronger predictors of popularity than emotional features. I produced statistical visualizations and quantitative interpretations to explain the findings.

The work combined data preparation, statistical modeling, diagnostic assessment, and communication of results. The cover is a newly prepared method overview; it is not a figure from the original academic report.