Inspiration
We were tired of being bombarded with the same TikTok viral sounds and wanted to discover new songs that actually suit our taste. With this in mind, we also wanted to make the app interesting for individuals like ourselves who are slowly losing their brain cells that help them focus.
What it does
Spot-A-Song gives personalized song recommendations based on your Spotify playlists by highlighting 30 seconds worth of ✨ vibes ✨ in a Tinder-like format. It uses Computer Vision to figure out if it's a slay or a nay. Data about user listening stats can be visualized in bubble chart, so that they can be more in tune with their listening habits.
How we built it
We used NodeJs and React for the backend and the frontend, respectively. The Spotify API was used to collect user data, while Google Cloud Vision API helped us with emotion recognition via facial features.
Challenges we ran into
Setting up Google Cloud Vision account permissions was a hassle, and attempting to deploy it was even worse :D Every feature we attempted to implement resulted in further bugs.
Accomplishments that we're proud of
The base features that we aimed to do were achieved. We weren't certain if we would actually be able to implement emotion recognition but like Dora said, we did it!! Lo hicimos <3
What we learned
Fix one bug at a time, Concordia food makes us gassy 💨, we work better with our thinking caps on and no eyebrows
What's next for Spot-A-Song
Computer Vision with gestures/other motion rather than faces, synced lyrics, filtering on genres and other fields, mobile version/responsiveness
Test out Spot-A-Song -> https://magenta-tarsier-db0bea.netlify.app
Frontend repo -> https://github.com/jfv-2000/spotasong-frontend/
Backend repo -> https://github.com/jfv-2000/spotasong-service/
Log in or sign up for Devpost to join the conversation.