Inspiration
Do you ever feel like the playlist you are listening to just doesn't fit your mood? Well, this was a pretty common problem for us, so we decided to make Spotifeeling.
What it does
Spotifeeling is a web application which plays music related to your mood. All you have to do is turn on your camera - the app will detect your mood and play a song which it feels will match your mood.
How we built it
We first learned about the Spotify API, which we used to collect data on songs. With this data, we used a machine learning model to classify songs into different moods. Next, we used Face API to both recognize facial expressions and then categorize the facial expressions into different moods. These moods were then connected to the moods detected from the songs in the playlist provided by the user to give songs which matched people's mood.
Challenges we ran into
We had problems merging our python backend with node.js.
Accomplishments that we're proud of
It was very nice to see the portion which categorizes songs by mood work (we spent several minutes just playing around to see how it categorized popular songs). We were also proud when the mood classifier of the facial recognition portion of the app work properly.
What we learned
We learned about the Spotify API to access and retrieve data about songs, along with making a widget on our website. To work with the facial recognition models, we learned about Face API.
What's next for Spotifeeling
- Connect calendars: Are you planning on reading a book from 1-3 and studying from 4-6? Spotifeeling will have the perfect music for you! We would want to connect Spotifeeling to your google calendar so that it is updated on whichever music might suit you at that moment.
- Add more moods: Feel tired while working on a boring assignment? In the future, we want to add more moods such as “tired” so that Spotifeeling can tell when you feel sleepy to play music that wakes you up!

Log in or sign up for Devpost to join the conversation.