Generating Classical Music Playlists using Radio Broadcast Programming

Recommender systems are algorithms that aim to suggest relevant items to users, such as movies to watch, products to buy, or in the case of our interest, music to listen to. In recent decades, with the rise of many web services, they have increasingly taken hold in our lives. Current music recommendation systems trace back, historically, to the work of music programmers and disk jockeys in commercial and institutional radios in the XX century. These insiders ranged from self-taught youth to highly distinguished musicologists and music critics and provided an extreme diversity of musical propositions. Of course, their proposals were all indistinctly biased because they were the product of human minds and knowledge. And of course, many (but not all) of such biases were driven by explicit and/or implicit record labels’ strong thrusts in the form of bribes, money, benefits etc. However, this scenario also provided a small cohort of culturally and politically motivated people who were committed to some form of cultural enhancement. This allowed for a small niche diversity in the general picture