Recommending Music Titles

Consumer choices and how they can be influenced are critical factors for every business. For instance, most people are interested in specific music genres, have favorite authors, or engage in particular hobbies. This information can be extracted from their purchase history or product reviews, and when utilized correctly, it can drastically increase the company’s profit. A frequently cited case is the one million dollar prize awarded by Netflix in 2009 to a team that developed an algorithm that increased the accuracy of the company’s recommendation engine by 10%. In the end, as more user interactions occur on any online platform, more data is available for analysis, leading to superior customized recommendations.

This chapter seeks to exploit product and user data to create recommender systems for music titles. We will base the discussion on a corpus of customer reviews from the Amazon online store. First, we will perform exploratory data analysis...