Using the Spotify API, incorporated audio features to explore how to make a better, content-based music recommender. Using genre-specific Spotify playlists and genre-tagged songs from the Million Songs Dataset, built a gradient boosted trees model with XGBoost to classify songs using audio content features. Insights drawn from the project concluded that given the current state of mostly content-agnostic recommendations, incorporating audio content could be a compelling future approach to recommending music.