seeds4tracks
Recently I have been reviewing old projects and ideas I've had, basically going over my portfolio.
I stumbled across an old project that I coined seeds4tracks.
The basic premise of this project being that you input "seeds", those being artists, tracks, and such. Then these "seeds" are used to generate a set of tracks that hopefully introduces you to some new music to enjoy.
The project didn't have a GitHub repository or anything set up for it yet, and it was just sitting on my machine, but was completed to a good point of functionality. One caveat being "seeds" can only be artists right now.
I decided to breathe some life back into this project and get it back up and running.
With that, I'm happy to say that seeds4tracks.com is live!
A repository backing it is also now available:
github.com/rcasto/seeds4tracks
Give it a shot, and hopefully find some new jams!
This project revival is of course very early, but what's there now is at least the core functionality, making use of the Spotify Web API.
Luckily, the Spotify Web API already has a recommendations endpoint. And it nicely does not require any user authentication, meaning users don't need to be logged into Spotify itself.
The only limitation with this API is that only a maximum of 5 "seed" objects total are able to be sent in a single request.
To get around this, I generate unique combinations between the input artists of groups of 5, then from these randomly select a certain number of sets to utilize to retrieve recommendation tracks. The results from these requests are aggregated, then de-duped, and then recommendation tracks are randomly picked.