🎬 What to Watch? – AI Movie Recommendation System

A Python + Flask based recommendation engine using IMDb datasets, offering fast search, filtered results, and personalized movie suggestions.

📌 Project Overview

This project is an AI-powered movie recommendation platform that provides:

• 🔍 Movie Search (title-based)
• 🎭 Genre-based recommendations
• 🌍 Region & language filtering
• 🌟 Personalized recommendations based on user history
• ⚡ Instant loading using a preprocessed preview dataset
• 🖥️ Modern frontend built with HTML, CSS, JavaScript
• 🔗 Backend API built using Flask

The platform uses IMDb datasets (basics.tsv, ratings.tsv, akas.tsv) to generate a lightweight preview JSON file for fast loading and high performance.

---

🚀 Features

1. Instant Movie Loading

IMDb dataset contains over 600,000 movies, but loading this directly takes minutes.
We use a preview file (movies_preview.json) built using:

• Movies after 1990
• Movies with > 10,000 votes
• Popular movies sorted by rating

This reduces load time to less than 1 second.

2. Search Engine

Search movies by title or genre.
Top results appear based on IMDb votes and rating.

3. Filtered Recommendations

Filter movies by:

• Genre
• Region
• Language

The backend returns the best-rated movies from the preview dataset.

4. Personalized Recommendations

The system tracks user history in localStorage:

• Genre history
• Region history
• Language history

Then generates dynamic personalized results.

---

🏗️ Tech Stack

Frontend

• HTML
• CSS
• JavaScript

Backend

• Python
• Flask
• IMDb official datasets
• JSON preview builder

Tools

• Git / GitHub
• VS Code
• pip / virtual environment

---

📂 Project Structure

project/
│
├── app.py
├── movies_preview.json
├── build_preview_modern.py
│
├── imdb/                 # (ignored in GitHub)
│   ├── title.basics.tsv.gz
│   ├── title.ratings.tsv.gz
│   └── title.akas.tsv.gz
│
├── public/
│   ├── index.html
│   ├── script.js
│   └── style.css
│
├── venv/                 # local environment
└── README.md

---

🔧 Installation & Setup

1. Clone the repository

git clone https://github.com/ErrorAayushy/recommen_engin.git
cd recommen_engin

2. Create virtual environment

python -m venv venv
venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Run the server

python app.py

Your browser will open automatically at:

http://localhost:5000

---

📘 How Preview Dataset Works

Instead of loading 600k+ IMDb movies during runtime, we preprocess them using:

python build_preview_modern.py

This generates:

movies_preview.json

This file is loaded instantly on server startup.

---

🧪 API Endpoints

Endpoint	Method	Description
/api/status	GET	Checks if movie data is loaded
/api/progress	GET	Loading progress (ETA, %)
/api/search?q=	GET	Search for movies
/api/recommend/genre	GET	Recommendations with filters
/api/recommend/personal	POST	Personalized suggestions

---

🎯 Future Enhancements

• Integrate TMDB API for movie posters
• Deploy backend online (Render, Railway, Vercel)
• Add user login + watchlist
• Add AI embeddings model for similarity (BERT, ANN)
• Mobile application interface

---

👨‍💻 Team Members

Name	Role
Aayushy S	Backend & Integration
Yashvi Shah	Collaboration Filtering
Princy Thakkar	Content-Based Filtering
Sonu Thakur	Backend Engineer
Aanchal Thosar	Dataset Engineering

---

📜 License

This project is for educational & academic purposes.

---

⭐ Acknowledgments

IMDb for providing open datasets.
Flask and Python community.
Team members who contributed to development.