This tutorial code uses the classic Scotch comparison dataset (foudn in Data Science for Business by O'Reilly Media) to build a recommendation system. We first build a cosine similarity matrix for our recommendations using Python Pandas, output the results to a CSV, create a sample Flask application to serve results, and finally Dockerize the application for deployment.
This tutorial was originally built for the SlimDevOps Twitch stream. More information can be found here.