Recommender system for books

Question

Does anybody know good recommender system for books. Lets say we have 20.000 books and 60.000 customers and 200.000 orders in last 10 years with 400.000 books purchased in diffrent generes.

Which data from books, customers and orders should we use to recommend best next book for specific customer to read?

Any good working model that you know?

Our top contibutor would sure like to read more Python Data Science books ,=)

Happy 2019 ,)

Answer 1

By the way, you can often find research datasets + code supporting ML projects.

Here's a research "Book Crossing" dataset for ~1M ratings for ~270K books.
http://www2.informatik.uni-freiburg.de/~cziegler/BX/

And an article using it with some Python links/codes:
https://towardsdatascience.com/my-journey-to-building-book-recommendation-system-5ec959c41847

Answer 2

"People Who Bought This:X May Also Like..." is a classical machine learning problem.

[just hinting, because I haven't looked for updates recently]
Andrew Ng's MLClass is really old at this point but it contains exactly that sort of recommender for movies.
I'd still take the course for the fundamentals, but better stuff might be out there.