🎬 Recommender Systems Collection

This repository contains two distinct recommender system projects:

1. Recommender-Systems — a simple movie recommender based on user–movie rating matrices.
2. Movie Recommender Website — a content-based movie recommendation web app built using Streamlit and Python.

---

🧩 1. Recommender-Systems

Overview

A foundational recommender system implementing collaborative filtering using a movie–user ratings matrix.
It predicts user preferences based on similarities between items (movies) and/or users.

Core Concepts

• Input: Movie–User ratings matrix (e.g., ratings.csv from MovieLens dataset).
• Techniques:
  • Item-based collaborative filtering using cosine similarity.
  • Matrix factorization (SVD-based) for dimensionality reduction.
• Output: Top‑N movie recommendations per user.

Workflow

1. Preprocess the movie–user rating dataset.
2. Compute a similarity matrix between movies or users.
3. For a given user, find the top‑rated movies.
4. Recommend similar movies that the user hasn’t seen yet.

Quick Start

• Run recommender_system_latent_factor.ipynb using Jupyter Notebook.

🌐 2. Movie Recommender Website

Brief:

A Streamlit‑powered web app that recommends similar movies to any movie the user selects. It uses a content‑based filtering approach derived from text tag similarity.

Tech Stack:

• Frontend: Streamlit (run from PyCharm)
• Backend: Python

Features:

• Frontend: Streamlit (runs locally from PyCharm or terminal)
• Backend: Python
• ML Libraries: scikit‑learn, pandas, numpy, requests
• Dataset: TMDB Movie Metadata

ML Model

1. Combine key text features (genres, cast, director, etc.) into a single tags column.
2. Vectorize tags using CountVectorizer (max features = 5000).
3. Compute cosine similarity between all movie vectors.
4. Return top‑N most similar movies to the selected one.

Dataset

https://www.kaggle.com/datasets/tmdb/tmdb-movie-metadata

Instructions to Run

• Run movie_recommender_system.ipynb using Jupyter Notebook.
• 3 new files should be added into the folder: movies.pkl, movie_dict.pkl, similarity.pkl.
• Add these files into the movie-rec folder.
• Go to https://www.themoviedb.org/ and sign up for a new account.
• Go to settings -> API -> Generate API and generate a new API key.
• Open the movie-rec folder in PyCharm.
• In the movie-rec folder:
  • Copy the APIv3 key and paste it in a .env file in the movie-rec folder.
  • The .env file should look like this:

  API=<api_key>
  

  • Open the terminal in PyCharm.
  • Run $pip install -r requirements.txt
  • Run streamlit run app.py

Website Preview

👩‍💻 Author

Developed by Shreya Bollimuntha