CS221 Project Final Report Community Detection based on Music Listening Habits

Music plays a big role in our lives and forms a billion dollar industry. A lot of research goes into music recommendation and prediction especially with so many internet radio services coming up like Spotify, Pandora etc. Even though community detection, in general, is a well tackled problem in social and information flow networks, an interesting and fairly unsolved problem is that of identifying communities within the music listening population. We want to tackle this problem as it can provide various insights into listening habits of people. As this is a unsupervised learning problem, we also build a recommendation system to suggest songs to our users based on the clustering. This helps us in evaluating our clustering model.