Spotify Recommender System - Artists

When you go Spotify app, you will see multiple recommendations done by Spotify like Songs, Playlists, Albums, and Artists. In this exercise, we will be building an Artist recommender system that will recommend Artists to listeners using Collaborative Filtering and Alternating Least Squares.

We will be using the LAST.FM dataset that was collected in May 2011 by Ignacio Fernandez-Tobias with the collaboration of Ivan Cantador and Alejandro Bellogin, members of the Information Retrieval group at Universidad Autonoma de Madrid (http://ir.ii.uam.es)

a. Fundamentals

There are three main recommendation methods used in streaming services:

• Popularity Based: recommends songs (or can be a book, video, movie, etc.) based on their popularity

• Content Based Filtering: focuses on the content and uses all the information about songs in the form of metadata, description, or tags (i.e. mood tag, genre tag, etc), match the songs with the preference of users, and recommends the ones that are in line with users' tastes.

• Collaborative Filtering: focuses on users information, completely disregards music contents, clusters users based on their preferences, and recommends songs to users with respect to their peers' preferences whom they share the cluster with.

NOTE: I WILL BE ADDING REST OF THE CONTENT SOON!