Improve and expand user interaction and connectivity on Spotify


Inspiration for this project came from our personal desire and interaction with the Spotify App. Music is inherently social, and we believe that this graph solution improves people’s quality of life by enhancing their musical experience through increased interaction and the ability to find people with similar music tastes.

What it does

This project aims to provide a framework, built on TigerGraph, which Spotify can implement to improve the interaction between their customers by determining how similar a user’s music taste is to their friends/other users. Potential features which Spotify could implement; user similarity scores, see common artists/genres between users, discord/subreddit style channels for people to share new music and interact, public user following recommendations. These features could be implemented by querying the network of users which has been built in TigerGraph.
How we built it
In this project, example user network data has been extracted from Spotify using the Spotify API and loaded into a TigerGraph instance. To determine the similarity of user’s music a query can be run which calculates the number of shortest paths (either length 2 or 4) between users, where the greater the number of paths, the more similar their music taste. In addition to this, a user’s song attributes (e.g., danceability, tempo, valance etc.) can be aggregated in order to determine what “type” of music they listen to. The design of our TigerGraph schema was focused on setting up nodes based on aspects we wanted to filter on. For example, we wanted to be able to filter track connections between users based on certain features of the track, and therefore, we split the features of tracks into separate nodes. Each edge then contains the elements that connect its nodes.