I maintain Movie-Recommendation-System, a Jupyter Notebook repository on @melonmusk20’s GitHub. It is a Streamlit app for discovering movies similar to a selected favorite, with posters fetched from TMDB. My repository listing is also recorded on github profile.
My Movie-Recommendation-System project started 2025-08 and was last updated 2026-11.
My system displays the top 10 recommendations based on a selected movie, along with posters.
My README identifies cosine similarity as the method for calculating similarity between movies and selecting recommendations.
My app uses Streamlit to provide an interactive interface for choosing a movie and viewing recommendations.
My app fetches movie posters using the TMDB API.
My README points to the TMDB movie-metadata dataset on Kaggle: https://www.kaggle.com/datasets/tmdb/tmdb-movie-metadata.
My processed movie data is stored in movie_data.pkl.
My repository’s movie_data.pkl is only a link placeholder because the file is too large for GitHub; I must download the real file from the Drive link and replace it.
pip install -r requirements.txt
export TMDB_API_KEY=your_key
TMDB_API_KEY = "your_key"
streamlit run app.py
I open http://localhost:8501, select a movie from the dropdown, and click “Recommend” to see the recommendations and posters.
My README is inconsistent about what the similarity calculation compares: it describes analyzing movie features, but also says cosine similarity measures similarity between movie titles. It does not identify the exact representation or preprocessing, explain matrix construction or tie handling, or say whether the selected movie is excluded. It reports no quantitative evaluation or recommendation-quality results.
My README image declares dimensions of 595 × 842 pixels.
My project uses machine learning techniques to analyze movie features and find similarities between movies.
My dataset contains information about movie titles and IDs.
My README says cosine similarity measures similarity between movie titles.
My README provides this Google Drive link for the PKL file: https://drive.google.com/file/d/1wcVhYPEsU1iTp4ZUmjQZW2T9EAbleMyB/view?usp=sharing.