Python & Letterboxd — Build a Content Filtering Rec Sys From Scratch
This article refers to a project I did it with the movie reviews of the great site Letterboxd. The result is a Content Based Filtering Recommender System of current popular movies based on the tastes of a specific user. If you are interested in recommender systems, I wrote an article on the subject here.
First of all, like many people, I have a hard time finding movies that I might like when I’m ready to watch something. That’s because even if the movie looks great and would be perfect for me, I’m constantly thinking about finding a better one and keep scrolling through movies until I don’t want to watch anything else. With all the streaming platforms offering choices and our attention span shrinking, what should be a good time is turning into a waste of time.
This is essentially why any website providing products or services uses recommendations to encourage the user to consume or buy.
Recommender systems are, however, hard to build because of the data and computing power it needs to be efficient. Before diving in my project, I advise you to have a quick check about my article on recommender systems as I will only focus on how I built mine.
