Why we built it
We wanted to see if we could build a music matcher without relying on black-box recommendation APIs. We set out to write our own scoring logic that matches user surveys against a local track database.
2024Technical Leadunavailable
GrooveEstrella
A music matching app built in a team of five using a deterministic scoring matrix.
Instead of plugging in Spotify or Last.fm recommendation APIs, we wanted to build a music recommendation engine ourselves. We designed our own matching logic using survey responses and weighted preferences.
Availability Notice
Spotify now requires a Spotify Premium account for third-party Web Playback & API access to function. You can still visit and test the live application.
Case study
The Challenge
Qualitative preferences like mood or tempo are hard to map. We designed a simple scoring matrix. When a user selects a mood, our script weights matching genre tags and filters out tracks below a specific score threshold.
My Role
I coordinated the development workflow for our team of five and wrote the React frontend. I also helped write the scoring parser. This was my first experience managing git branches and pull requests across a team.
What I Learned
We wasted three days at the start because our database parser expected one date format while the frontend sent another. This taught me that before writing any code, team developers must agree on data structures first.
Next step
Let's talk about building products.
I am always glad to talk with other developers, founders, or recruitment teams who value clear communication and straightforward code.