beatSync - a knn based music recommender

beatSync is a knn ( a machine learning algorithm ) based music recommender. The dataset has ~3500 songs and recommends songs similiar songs based on the songs features such as dancebility, energy, popularity, etc.

Demo

Fully working Live Demo - Click here
(First request takes longer times, after it response time is ~500ms - 800ms)

Tech Stack

Frontend

• Fronted is built on ReactJS with vanilla JS.
• Deployed on Vercel.

Backend

• Backend is built on Flask to serve the model,
• Deployed on Render (First request will take a longer time, render free-tier. Normal requests take 500ms - 800ms)

Model

• Used scikit-learn, pandas.
• sickit-learn knn.neighbours for implementing KNN.

Dataset

• Datset is in the repo.
• has ~3500 songs with 13 features each.