This is a content based movie recommender system that makes suggestions based on the similarity between movie plots, genres, cast etc.
The project follows the below steps:
- Load and pre-process the TMDB movie metadata dataset
- Combine important features into a tags column
- Vectorize tags into TF-IDF vectors
- Calculate cosine similarity between movie vectors
- Make recommendations by finding most similar movies
The core components are:
data_preprocessing.ipynb: Data cleaning and feature engineeringmodel_building.ipynb: TF-IDF vectorization and similarity calculationapp.py: Streamlit app for movie selection and recommendationsutils.py: Contains helper functions
To run the movie recommender app:
streamlit run app.py
This will launch the webapp where you can pick a movie title and get recommendations.
The movie similarity matrices are saved as pickle files to avoid recalculation.
Some ways to improve the project in future:
- Use better similarity metrics like Jaccard or fuzzy matching
- Incorporate ratings, watched status for personalization
- Add additional metadata like actors, directors etc.
- Use image based similarity to find visually similar movies
- Deploy the app on a cloud platform for increased availability