Have you ever felt overwhelmed when choosing a song from Spotify?
This is the moment when ISP will help you.
All you need to do is take a selfie. ISP will do the rest of the magic* and provide you with a song that fits your musical taste, feeling and environment.
Don't worry, the picture stays on your device.
*Machine learning for face recognition, emotion classification, fancy song search algorithms and other lovely stuff
💥 Try it out right now at intelligent-song-picker.com!
I developed this project during the course Interaction/Programming by Prof. Dr. Thomas Rist at the University of Applied Sciences Augsburg. In addition to the web application I wrote a paper about this project.
Create a local running copy of this project
cd client yarn install
Start local development server
Or create a build for deployment in production
cd server yarn install
Run local nodemon server with hotreloading
Or run production server on node only
This is a list of the core dependencies used in this project for your further reference.
- tfjs-emotion-classification (not a "real" dependecy but implemented)
- Improve stability
- Add Autoplay function to go on playing songs after the first one is finished
- Handle Spotify non-premium users
- Add contenthash to js files for improved caching
- Create robust mobile version (currently a very bad workaround is in place here)
- Reduce own queries to spotify and instead leverage the capabilities of the Spotify Web Playback SDK