Movie Recommender System 🎬🍿

Project Overview

• This is a Movie Recommender System built using Python, Streamlit, and machine learning techniques. The project is a result of self-study and hands-on learning in Data Science and Machine Learning. It allows users to select a movie and get 5 personalized movie recommendations with posters fetched from TMDB API.

• The app demonstrates the entire workflow from data preprocessing, feature extraction, similarity computation, to deployment, making it a full-stack Machine Learning project.

Features

• Search and select a movie from a large dataset.
• Recommend 5 similar movies based on content similarity.
• Display movie posters alongside recommendations.
• Uses a similarity matrix for fast lookup.
• Handles large datasets using Google Drive integration for heavy .pkl files.
• Deployed live using Streamlit Community Cloud.

Technologies & Packages Used

• Python – Core programming language
• Streamlit – Web app framework
• Pandas – Data manipulation
• Pickle – Save/load serialized Python objects
• Requests – API calls to TMDB for movie posters
• gdown – Download large files from Google Drive
• NumPy – Numerical computing
• scikit-learn – Machine learning & vectorization

Data Preprocessing & Feature Engineering

• Data Collection: Combined two data sources into a single DataFrame.
• Data Cleaning: Removed duplicates, handled missing values, standardized formats.
• Feature Selection: Selected relevant columns for recommendations (title, genres, overview, cast, crew, keywords).
• Vectorization: Used CountVectorizer / TF-IDF to convert text features into numerical vectors.
• Similarity Computation: Calculated cosine similarity between movie vectors to find the closest matches.

Deployment

• Platform: Streamlit Community Cloud (free)

File Structure

Movie-Recommendation-App/
 ├── app.py                # Main Streamlit app
 ├── Assets/               # Local data files
 │    └── movies.pkl       # Movies dataset
 ├── .streamlit/
 │    └── secrets.toml     # API keys & file IDs (not pushed to GitHub)
 ├── requirements.txt      # Required Python packages
 └── README.md             # Project description

How to Run Locally

Clone the repo:

git clone https://github.com/YourUsername/Movie-Recommendation-App.git
cd Movie-Recommendation-App

Install dependencies:

pip install -r requirements.txt

Add .streamlit/secrets.toml:

API_KEY = "your_tmdb_api_key_here"
FILE_ID = "your_google_drive_file_id_here"

Run the app:

streamlit run app.py

Live Demo

Click here to see the live app on Streamlit

GitHub Repository

View the GitHub repo

Author

• Dilan Liyanaarachchi

YouTube Demo

Watch the YouTube Demo