Over the weekend, I had the absolute pleasure of attending the RythmHacks 2025 hackathon in Waterloo, Ontario. A two day event at the Accelerator Centre where students learn through hands-on workshops and build a project, start-to-finish with a demo at the end and prizes to be won. If it sounds exhausting, that’s because it is.
On the first day of the event I lead a workshop where I talked about how recommendation algorithms work. We looked at how music is modeled using data and how the algorithm can suggest a next song for you and get consistently more accurate with what you like. For the event, I build a web app that demostrates content-based filtering.
Try the demo
👉 https://adamstirtan.github.io/rhythmhacks2025
How it works
Rate a song using the 1-5 stars and notice that the weights for each musical dimension moves. The next song will be selected based on the highest matched, unplayed song. For more information on the workshop presentation, check out the slides.