nksaraf / song-discovery

K-means clustering based song recommendation system

Song Discovery

Cornell CS 3300 Final Project (done with Michael Gingras and Kat Bau)

Web App Link: https://songdiscovery582.firebaseapp.com/

A simple song recommendation system based on a user's Spotify playlists used to demonstrate and visualize the k-means clustering algorithm. The Spotify API is used to get songs' characteristics like danceability, energy and acousticness as well as allow users to import playlists. The website uses the user's playlist to match songs to a stock set of songs with a wide range of characteristics. It then recommends songs from the cluster that has the maximum number of songs from user's playlist.

Technologies used: HTML, CSS, JavaScript, D3.js, Spotify API, Firebase