🎬 Movie Recommender System

A content-based Movie Recommendation System built using Python, Flask, and machine learning techniques.
The application recommends similar movies based on user selection and dynamically fetches movie posters using the TMDB API.

---

📊 Dataset

This project uses the TMDB Movie Metadata Dataset:

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

The CSV dataset files are included inside the dataset/ directory.

Generated model artifacts (.pkl files) are not stored in this repository due to GitHub file size limitations and must be created locally. pkl files can be genereated by running model_builder.py. They are generated in the dataset folder.

---

📸 Demo Screenshots

Home Page

Recommendation Results

⚙️ Installation

• Clone the repository:

    git clone https://github.com/uraza116/Movie-recommender-system.git
    cd Movie-recommender-system


python -m venv .venv

- Activate:

    Windows:
    .venv\Scripts\activate

    Mac/Linux:
    source .venv/bin/activate

- Install dependencies:

    pip install -r requirements.txt

- Generated model artifacts (`.pkl` files) are not stored in this repository due to GitHub file size limitations and must be created locally.
pkl files can be genereated by running model_builder.py. They are generated in the dataset folder. Run following command to generate those files

    python .\model_builder.py

-Run app.py to run flask app
    python.exe .\app.py

-Enter following address on web browser after running the Flask app
    http://127.0.0.1:5000/

🚀 Features

• Content-based recommendation system
• TF-IDF vectorization of movie metadata
• Cosine similarity computation
• Precomputed similarity matrix
• Dynamic movie poster retrieval using TMDB API
• Interactive Flask web interface

---

🛠️ Tech Stack

• Python
• Flask
• Pandas
• NumPy
• Scikit-learn
• HTML / CSS

---

🧠 Technical Approach

1️⃣ Data Processing

• Merge relevant metadata columns
• Clean and preprocess text fields
• Combine features into a single content feature

2️⃣ Feature Engineering

• Apply TF-IDF Vectorization
• Convert textual movie descriptions into numerical vectors

3️⃣ Similarity Computation

• Compute Cosine Similarity
• Generate similarity matrix between all movies
• Store results in similarity.pkl

4️⃣ Recommendation Logic

When a user selects a movie:

• Locate movie index
• Retrieve similarity scores
• Sort and select top-N similar movies
• Fetch posters via TMDB API
• Return recommendations to frontend

📂 Project Structure

Movie-recommender-system/ │ ├── dataset/ # TMDB CSV files ├── static/ │ └── css/ ├── templates/ │ └── index.html ├── app.py # Flask application ├── model_builder.py # Generates movies.pkl and similarity.pkl ├── movie.ipynb # Development & experimentation notebook ├── requirements.txt ├── .gitignore └── README.md

---

📈 Generate Model Files

• Before running the application, generate the required model artifacts:

• python model_builder.py

• This will create: movies.pkl & similarity.pkl

• These files are required for the Flask application to run.

▶️ Run the Application python app.py

Open in your browser: http://127.0.0.1:5000

🔑 TMDB API Setup

• This project uses The Movie Database (TMDB) API to fetch movie posters.
• Create an account at https://www.themoviedb.org
• Generate an API key
• Replace the API key inside app.py

👨‍💻 Author

Usman Raza PhD Candidate | Backend & ML Engineer