Movie Recommendation System - Find Similar Movies Using LLM
Background (The Story)
How often do you wonder what to watch?π€especially on weekends!
There are so many good movies to pick from!πΏ
I got you covered! π Find similar movies quick and easy! With LLM we get results which are not just relavant but which are very similar in context of stories, genre, cast, makers and more!
So let's stop wondering and start watching!
Explore π
Wondering what to watch? you can go to home page to explore popular movies and find similar movies based on your selection!
Search π
You can search for ANYTHING!
Movies like inception
Love story of writers
Movies of tom cruise
It will give you results based on the context of your search, similarity in plot, genre, makers, cast and summaries of 5000+ popular movies.
Watchlist π
If you wish to save some movies for later, you can do that by adding them to your watchlist i.e. My Watchlist
You can add / remove as many as you like! π
Enough talk! Let's pick and watch a movie! πΏ
https://wondering-what-to-watch.vercel.app/
Behind the scenes π₯π€
Gethering movies data π
Vectorizing summaries for similarity search βοΈ
Building frontend β¨
Gethering data π
I used themoviedb.org to get a list of popular movies. I then used this list to make a CSV file of movie details. The CSV file has the following info: title, rating, genre, cast, crew, directors, producers, overview etc
Below is python3 code for fetching movie details and storing into csv file for later use.
Hosting and Vectorizing Data βοΈ
This project uses Astra DB to store the data. The text-embedding-3-large model, which is trained on a massive corpus of text, is used to vectorize movie summaries. This model encodes text into vectors that can be compared to find similar movies.
Building Frontend π
I have used NextJs along with AceternityUI for building the application.
Below is the source code the project
π https://github.com/tapanrachchh/wondering-what-to-watch
Thanks for reading this article, Hope you liked my idea, Let me know if you have any feedback or idea! to make it more interesting!βοΈπ
Subscribe to my newsletter
Read articles from Tapan Rachchh directly inside your inbox. Subscribe to the newsletter, and don't miss out.
Written by