As you may know, watching (and reviewing!) movies is one of my favorite things to do. A while back I had made a local movie recommender script that I had been using every now and then when I wanted to find a new movie to watch. There is a writeup of that script in my earlier article, but the idea is simple: in order to increase your odds of finding a movie that you might like, find people that like the same movies as you and then find the movies that you like that I haven’t seen. So there are two datasets in play: my movie reviews, and the reviews of millions of other people.
Anyway, I finally got around to making this a web app and putting it into a container so that I can self host it. I look forward to converting more of my local data science projects into web apps now that I have a cloudflare tunnel running into my host, which makes hosting (checks notes) free. I had wanted to allow people to point to their own set of reviews (such as at letterboxd.com) but I am still waiting for my api key to be able to pull back the reviews for an individual user).
So with that, here it is: recommendmovies.adambytes.net. You might be prompted to authenticate with a gmail account – I am working to set up anonymous access.
You can change some of the parameters to get different results.
The movie posters and description are then displayed.
The genres for the current result set are then displayed for easy filtering.
And finally, movies can be filtered based on year (which comes in handy, given my preference for newer movies).
