Here I explore online sales data for a wine store. In part I, I use k-means clustering to segment the customers. In Part II, I use matrix factorization to build a wine recommender.
recommenders.py has functions for building recommenders.
WineDrinkers.ipynb goes through the workflow of clustering wine drinkers.
WineDrinkers2.ipynb goes through the workflow of building wine recommenders.
To see a full description, see the blog post here.