I used the API of a popular music website to scrape the listening behavior of its users. These data tell me which albums each user listened to. I used those data to build a recommendation system!
These recommendations were built using the Apriori algorithm for item recommendation. The data for the model was collected in mid-May 2020. This algorithm is vulnerable to the cold start problem, so it will have poor results for albums released very close to when I scraped the data.
You can search for music below. The search runs on a generic whoosh index, and I did not put a great deal of effort into making it good. Sorry!