The Utilization of Content Based Filtering for Spotify Music Recommendation

The purpose of a Recommender System is to provide relevant suggestions to a group of users for things or products that they may be interested in. Recommender systems were created to bridge that gap between information gathering and analysis by filtering all available data to offer only what is most important to the user. In this paper, we experiment to make a music recommendation based on the user's playlist. We use the content-based filtering specifically cosine similarity to make a recommendation system and try to compare the recommendation we have made with Spotify recommendation. Our research found that by using cosine similarity, the similarity result reaches up to 80% similarity for the song and 50% similarity for the artist which mean cosine similarity works well for our recommendation system.