Twitter bot that posts weekly top films and trends, replies with personalized movie recommendations, and more utilizing sites like Letterboxd and TMDb.
Tweets Letterboxd‘s (a social movie site) weekly top films and trends every Saturday at 5 PM and replies to mentions with personalized movie recommendations based on a user’s Letterboxd diary entries.
https://twitter.com/LetterBotFilm
Usage: Tweet to the bot account ‘@LetterBotFilm [Letterboxd_username]‘
Hosted on a Raspberry Pi 3.
Automates the twitter bot’s functionalities on a schedule and uses StreamListener to receive twitter messages in real time. The Tweepy library was used for accessing the Twitter API.
A web scraper to collect Letterboxd’s most popular films of the week and a user’s recent movie entries and wish list using the Selenium and BeautifulSoup libraries. The collected data is stored on a SQLite database.
A content-based movie recommender system based on the cosine similarities between vectorized film attributes from a generated user profile and TMDb datasets using the Pandas and Scikit-learn libraries.
Contains GET requests to TMDB API for movie metadata and poster.
Generates charts for weekly popular films and recommended lists to be tweeted out using the Matplotlib library.
Contains ‘tmdb_5000_credits.csv’ and ‘tmdb_5000_movies.csv’ files provided by The Movie Database (TMDb) and available on Kaggle. These datasets include movie metadata used by the recommender system.
Temporarily holds generated chart visuals in preperation to be tweeted.