Create Music Recommendation System Using Python

Original article can be found here (source): Artificial Intelligence on Medium

“Visualization gives you answers to questions you didn’t know you had.” — Ben Schneiderman

After visualizing the data in Tableau, it was observed that we have a lot of 0’s in the Year variable. But we didn’t remove those observations since we are not going to use them anyway. The above graphs are Tree Maps. It was also observed that Coldplay is the most listened Artist in the data set with 36,406 hits.

Business Problem

Too many choices can overwhelm users. If offered too many options, the user may not buy anything. Streaming services like Spotify have massive catalogs. Identifying the tracks a user might like and giving them a limited choice becomes an import task. Or else the user might leave the platform.

We are going to create to playlists, one being a Popular Playlist, and identifying songs that are similar to a song a user already likes — hence giving them personalized recommendations.

Recommendation Systems

Content-Based Recommendation Systems are used when you lack data about user preferences. You need labeled data to create this type of recommendation system. When a user watches a movie from the genre of Comedy, the system recommends other movies which are also from the same genre. That’s how a basic CBR Systems works. In this project, we are not building a CBR System since we already have labeled data.