The idea for this project came from my master's thesis, which focused on movie recommendation algorithms. Having implemented a command line system, which allowed the users to rate the movies, and receive recommendations, I thought it will be fun to do it in a more user-friendly way, in a form of a web application
- Python 3.x
- Django 3.x
- HTML
- CSS
- JavaScript
- AJAX
- NumPy
Inside mysite/static/ directory, create new folder, called 'private'. Inside this newly created folder, create a new file, called 'passes.json'. This file should be structured like: {"key":"<key>","email":"<email>":,"password":"<password>"}, where:
- <key> is Django's secret key
- <email> is your email address
- <password> is your email password (or your application password). (Unless you want to test the password reset functionality, <email> and <password> may contain completly bogus information).
Create and activate the virtual environment:
python -m venv venv
venv\Scripts\activate.bat
Install required dependencies, by running:
python install -r requirements.txt
- User authentication:
- Registration
- Logging in
- Password reset via email
- When the new user logs in for the first time, he is presented a set of movies to rate,
to establish his/her initial preferences:
- All movies are displayed, they can be sorted by each column, and filtered by decades
- Searching movies by the title:
- A user can view all of his/her rated movies, and sort them, or filter them by decade, rating, or date:
- A user can add, update, and remove rating for a movie:
- Movie overviews and posters are loaded from TMDb, using TMDb API
- A user can view information about his/her ratings:
- A user can view the list of all users, and search them by the username:
- A user is presented with a list of recommended movies along with the possibilities that he'll like each of them:
- Write more tests in Selenium.
- Enhance the recommender algorithm by using collaborative filltering.
- Run the recommendation in the background using Celery.
- More detailed user stats (display most similar users with links to their profiles).