Used Javascript, Node.js/Express.js, Python, and Google Datastore to create a Spotify playlist recommender + insights web application :)
Music taste is unique to everyone and is hard to describe in words - with this app, you can compare your music preferences with friends and enjoy some new tracks!
Click on the images below to watch the Youtube demos!
- Phoebe (UI)
- UI beginning stages: colour palette, basic design; more responsive data population by Stephanie
- Assisted looking into machine learning models in beginning stages
- Stephanie (Spotify API, frontend + backend)
- Spotify API calls: login authorization flow, access tokens, get user, user playlists, playlist and track objects, personalized user insights, search Spotify user
- Frontend: communicate with server using Javascript fetching, ie. send selected user playlists, populate DOM with playlists and data after calling API, responsive navigation + grid layout to display playlists, cookies, user login status, lots of rendering!
- Backend: Node.js server, handles user authorization, page rendering, GET/POST requests, and Spotify API calls
- Client-server communication: Javascript fetch, GET, POST, JSON formatting/parsing
- Jessica C (Datastore)
- Created Datastore entities for tracks (name, artists, audio_featrures, track_id, playlist_id)
- Implemented functions for inserting/retrieving to/from Datastore, ie. addTracks(), getTrackByPlaylist(), getTrackById()
- Jessica F (Recommender)
- Modified Spotify's recommendation function to fit our app purpose and functionality, ex. getTracks(), getTracksByPlaylistId(), getRecommendations()
- Ran tracks from Datastore playlists through recommendation function to output a final playlist of recommended songs starting from 'most recommended'
- Git clone the repository
- You might have to delete the
node_modulesfolder and runnpm install - Navigate to working directory/repository
- In Terminal/Powershell, run
node server.jsornpm run devStart(nodemon) - Application should be running on local host port 8888!