🎬 Movie Recommendation System

A Content-Based Movie Recommendation System built with Python, Scikit-learn, and Streamlit. The application recommends movies similar to the one selected by the user using Cosine Similarity, while fetching movie posters dynamically from the TMDB API.

---

🚀 Live Demo

🔗 Streamlit App: Movie Recommendation System

---

📸 Application Preview

🏠 Home Page

🎬 Recommendation Page

---

✨ Features

• 🎥 Search from 5000+ movies
• 🤖 Content-based movie recommendations
• ⭐ Returns the Top 5 similar movies
• 🖼️ Fetches movie posters using the TMDB API
• ⚡ Parallel poster loading for faster performance
• ☁️ Deployed using Streamlit Community Cloud
• 🤗 Model files hosted on Hugging Face

---

🛠️ Tech Stack

• Python
• Streamlit
• Pandas
• NumPy
• Scikit-learn
• NLTK
• Requests
• TMDB API
• Hugging Face Hub

---

📂 Project Structure

movie-recommendation-system/
│
├── app.py
├── movie_recommend_system.ipynb
├── requirements.txt
├── runtime.txt
├── README.md
├── .gitignore
└── .env (local only)

---

🧠 How It Works

The recommendation system follows these steps:

1. Load the TMDB Movies and Credits datasets.
2. Merge both datasets.
3. Select important features:
   • Genres
   • Keywords
   • Cast
   • Director
   • Overview
4. Clean and preprocess the data.
5. Apply text stemming using PorterStemmer.
6. Convert text into vectors using CountVectorizer.
7. Compute movie similarity using Cosine Similarity.
8. Recommend the top 5 similar movies.
9. Fetch movie posters dynamically using the TMDB API.

---

📊 Machine Learning Pipeline

TMDB Movies Dataset
          │
          ▼
Merge Movies & Credits
          │
          ▼
Feature Selection
(Genres, Cast, Director,
Keywords, Overview)
          │
          ▼
Text Preprocessing
          │
          ▼
Porter Stemming
          │
          ▼
CountVectorizer
(max_features = 5000)
          │
          ▼
Cosine Similarity Matrix
          │
          ▼
Recommendation Engine
          │
          ▼
Streamlit Web Application

---

📚 Dataset

• TMDB 5000 Movies Dataset
• TMDB 5000 Credits Dataset

---

⚙️ Installation

Clone the repository

git clone https://github.com/Akhilesh-Mogaveer/movie-recommendation-system.git

Move into the project directory

cd movie-recommendation-system

Create a virtual environment

Windows

python -m venv .venv
.venv\Scripts\activate

Linux / macOS

python3 -m venv .venv
source .venv/bin/activate

Install the dependencies

pip install -r requirements.txt

---

🔑 Environment Variables

Create a .env file in the project root.

TMDB_API_KEY=YOUR_TMDB_API_KEY

Get your free API key from:

https://developer.themoviedb.org/

---

▶️ Run the Application

streamlit run app.py

---

🤗 Model Files

The recommendation model files are hosted on Hugging Face and downloaded automatically when the application starts.

Repository:

https://huggingface.co/Akhileshh1/movie-recommendation-system

---

📦 Generated Files

The notebook generates:

• movies_dict.pkl
• similarity.pkl

These files are used by the Streamlit application for fast recommendations.

---

🎯 Future Improvements

• ⭐ Display IMDb/TMDB ratings
• 🎭 Genre filtering
• 📅 Release year filtering
• ❤️ Favorite movies feature
• 🔍 Search autocomplete
• 🌍 Multi-language support
• 🤖 Hybrid recommendation system
• 🎬 Movie trailers

---

👨‍💻 Author

Akhilesh

• 💼 LinkedIn: https://linkedin.com/in/akhilesh-1109ma
• 💻 GitHub: https://github.com/Akhilesh-Mogaveer

---

🙏 Acknowledgements

• TMDB for providing the movie dataset and API.
• Scikit-learn for machine learning utilities.
• Streamlit for rapid web application development.
• Hugging Face for hosting the model files.

---

⭐ Support

If you found this project useful, please consider giving it a ⭐ on GitHub!

It helps others discover the project and supports future improvements.