MOOSIC - Capstone Project

How did we come up with our project idea?

We were talking about how unsatisfied we are with the music recommendations we get from Spotify or other music streaming platforms. When we talked about why we don’t like the recommendations we figured out, that the same songs were proposed partly repeatedly or the tracks didn’t fit the mood we were in or the activity we were engaged in. By thinking of ways how to tackle these problems we came to the conclusion, that it would be nice to have some kind of user input, to make the track recommendations fit the user preferences. We then pinpointed the two features mood and genre to act as filters for better predictions.

Why did we go with these two features mood and genre?

The problem with music recommendations at present is that the recommended tracks often times don’t fit the listeners mood. For example Spotify uses a complex system of a variety of factors to personalize the listening experience for each user. These factors include the user's listening history, their favorite songs and artists, and the genres and playlists they listen to. The problem is, that mood isn’t tracked. And it’s the most human thing to change moods over the course of time - be it days or hours. So, when we are sad on one day for example and listen extensively to sad songs, the algorithm is bound to recommend these songs again to the user, even if the user isn’t in a sad mood any more.

But why is it so important, that music fits our mood?

Participants of a study¹, conducted by British psychologists, responded that they use music to express emotions and manage their mood. Music becomes a tool for us to manage our moods.
¹Source: Why do we listen to music? A uses and gratifications analysis by Adam J. Lonsdale and Adrian C. North, p.120. Published in: British Journal of Psychology (2011), p.102, pp.108–134