Inspiration

We came up with the idea when thinking about how in certain situations such as a party or a long drive with a friend, it is always challengin to find music that you both want to listen to. Either one person is dominating the music selection to the dismay of the other, or possibly you struggle to find music that you both will enjoy listening to.

What it does

That is where Unify comes along. All you do is enter some spotify usernames and select some playlists that you like and our webapp will do the rest. We are developing a custom algorithm to merge you and your friends music tastes to create a playlist that everyone will enjoy. We are not just going to add the same songs as in both your playlists and shuffle them either, we are developing a custom algortihm to analyze the genre/artists that comprise your music tastes and to find suitable other songs that fall within the same range as the other playlists. If you are into hard techno and your friend is into mumble rap, we will find some songs that satisfy both of your music inclinations.

How I built it

We initally built the backend server in node.js and express with a react front end. Due to some of the data structures we wanted to use for our algorithm and the computations that would take place, we wanted to use Java to handle the heavy stuff. We learnt the hard way that trying to implement java functions in an express server that we hope to deploy onto the cloud no less... not a fun road to go down. Therefore we decided to switch over to a flask python backend server since we could still use Python to do everything we wanted to do in Java.

Challenges I ran into

Remaking the entire backend server and all the api calls in python from node was not great but hey it was a good learning experience.

What's next for Unity For Spotify

We have a TON of ideas to run with about how to expand the functionality of the site. Spotify provides you with an extremely robust amount of information about songs/artists/genres that we plan to take full advantage of. We hope to be able to create playlists out of not just tghe relation of the artists and the genre but the actual sound metrics of the songs themselves such as tempo/instrumental vs lyrical etc etc etc.

Share this project:

Updates