Inspiration

We found that traditional music streaming platforms known for their great recommendation systems fail when the user neglects telling the app whether they enjoyed the previous song. If I just let a music streaming service pick a whole playlist for me and I never stopped listening to it, the platform can't accurately judge my level of enjoyment or dissatisfaction.

What it does

We employ facial and emotional recognition, provided by Microsoft's API and improved by our ML algorithm, to better curate music recommendation.

How we built it

We leveraged our team's diverse technical backgrounds by creating back end servers in several technologies and orchestrating them together using cloud storage and web requests. These included Flask servers in Python and Express servers in NodeJS. The multiple front end aspects of our project were done with ReactJS.

Challenges we ran into

We wanted to expand our platform's reach by integrating them with Bose's speakers. However, we didn't have enough time to fully take advantage of their APIs.

Accomplishments that we're proud of

What we learned

What's next for Smart Song Recommender

Share this project:

Updates