Tea Recommendations
As part of a class project, I built a tea recommender. The basic goal of this is to allow users/consumers to discover new teas without being locked in to a single brand or store. Imagine teabox that works across a whole bunch of tea vendors – local B&M, online, etc.
While this tea recommender doesn’t learn, it uses tf-idf to vectorize tasting notes, cosine similarity scores to find similar and different teas, and finally a basic aggregation algorithm to find the best tea for any tea lover.
Type in a few teas you like or dislike and get ready to be amazed.