Bindle wants to make a personalized newsletter and regularly send TOP books based on users’ interests.
A data-only approach would be to check which books users read, how long did it take them (maybe they were just skimming) and how did they rate books after finishing. This approach is the foundation of many recommendation algorithms (check out collaborative filtering, for example).
Good news for us: Bindle asks users to directly fill in their interests on the profile page. Let’s try to use this data to see how popular detective books are.
...