Spotify_song_recommender

This is a short project to build a song recommender based on numerical features of songs obtained from the Spotify API

Structure

Model creation

In TrainingTheModel, the recommendation model is trained on about 22k songs that are pulled from playlists on Spotify aiming for a balanced selection of songs in order to get the best recommendations.

Recommender

Actual recommendation based on user input, either returns a song of the hot 100 or from the dataframe the model was trained on.