This is a college project in which we used the spotify dataset from kaggle (now taken down) and used multiple clustering algorithms to find clusters among the songs in the dataset.
Using evaluation metrics, the algorithm which performed the best was then used to recommend new songs to the user given a few songs the user has already listened to.