Inspiration
The lack of good music at parties is criminal. Pretty rare to find the DJs that are not absorbed in their playlist as well. Someone had to do it, so we did it.
What it does
It looks at the movement of the crowd using computer vision and using Spotify's data-set of over 42000 songs we find the song that makes you groove by matching the energy of the room. That is why we are the groove genie.
How we built it
We used OpenCV to find the change in the movement, and we take the derivative of the movement over the period of the song. Change in movement is directly proportional to the energy level. From Spotify's data set, we download (no ads yay!) and play the music that best describes the way the audience is feeling.
To further optimize the song choice we keep track of the 40 songs that the audience responded to as expected. These are the songs that are reflective of superior data quality. So we choose songs that are the most similar to these songs after you have heard over 40 songs, while still matching the energy.
Challenges we ran into
Stopping the similarity checks from overriding the movement score currently being displayed by the audience. We fixed this by changing how much value we gave to similarity.
Accomplishments that we're proud of
Being able to detect the speed of and changes in movement using OpenCV, and the automatic player.
What we learned
Movement detection using computer vision, data processing with Spotify's database, and integrating Youtube into the project to play songs.
What's next for Groove Genie
A solo mode for Groove Genie, for listening to music on your own, which changes the music based on your emotional reaction.
Log in or sign up for Devpost to join the conversation.