Sortify is an application that uses unsupervised machine learning to divide a large Spotify playlist into smaller playlists grouped by musical characteristics. Our team won Best Design at Northwestern's annual IEEE Showcase (hosted over Zoom due to Covid).

Sortify uses Spotify's Web API to access the user's Spotify account. Once authorized, Sortify uses a refreshable access token to make API requests. Both the front end and back-end are written in Python.

There are three main steps to the machine learning implementation:Sortify determines song similarity by looking at audio features provided by the Web API. Example features include "danceability" and "acousticness." The front-end generates an interactive Matplotlib to display playlist clusters.

I was responsible for the Web API authorization flow and the Tkinter front end. Challenges our team overcame included playlist name conflicts, sparse clusters, and multi-threading.