This project analyzes Spotify audio features to predict song likeability using logistic regression. Key features like danceability and valence, along with their interactions, were studied to improve prediction accuracy. The final model achieved 89.74% accuracy, offering insights into music preferences and audio feature dynamics.