GitHunt
NU

NusratBegum/Netflix-Recommendation-System

Netflix Data Analysis and Recommendation Algorithm

cosine-similaritydata-analysisdata-sciencedata-visualizationjupyter-notebookmachine-learningmachine-learning-algorithmsnetflixnlppythonrecommendation-systemrecommender-systemtf-idftf-idf-vectorizationtf-idf-vectorizervector

Netflix Movies & TV Shows - Recommendation System

Python
License
Jupyter

A comprehensive data analysis and content-based recommendation system for Netflix Movies and TV Shows using Python. This project performs extensive Exploratory Data Analysis (EDA) and builds an intelligent recommendation engine using machine learning techniques.

Table of Contents

Overview

This project analyzes Netflix's content library to uncover insights about viewing patterns, content distribution, and trends. It includes a sophisticated content-based recommendation system that suggests similar movies and TV shows based on various features like genres, cast, director, and descriptions.

What Makes This Project Special?

  • Comprehensive EDA: Deep dive into Netflix's content with 20+ visualizations
  • Advanced Feature Engineering: 14+ engineered features from raw data
  • Statistical Analysis: Hypothesis testing to validate insights
  • Smart Recommendations: TF-IDF and cosine similarity-based recommendation engine
  • Professional Visualizations: Netflix-themed color palette and styling
  • Interactive Analysis: Ready-to-use Jupyter notebook with detailed explanations

Features

  • Exploratory Data Analysis

    • Content distribution (Movies vs TV Shows)
    • Geographic analysis (content by country)
    • Temporal trends (release years, addition patterns)
    • Genre and rating analysis
    • Duration analysis

  • Statistical Testing

    • Release year distribution analysis
    • Content type vs rating association
    • Geographic content representation
    • Temporal trend analysis

  • Recommendation System

    • Content-based filtering
    • TF-IDF vectorization (5000 features)
    • Cosine similarity matching
    • Multi-feature recommendations (genre, cast, director, description)

  • Professional Visualizations

    • Netflix-branded color scheme
    • Interactive plots
    • Clean, publication-ready figures

Dataset

Source: Netflix Movies and TV Shows Dataset on Kaggle

Details:

  • 8,807 titles (Movies and TV Shows)
  • 12 features: show_id, type, title, director, cast, country, date_added, release_year, rating, duration, listed_in, description
  • Date Range: Content added from 2008 to 2021
  • Coverage: Global content from 100+ countries

Installation

Prerequisites

  • Python 3.8 or higher
  • Jupyter Notebook or JupyterLab
  • Kaggle account (for dataset download)

Setup

  1. Clone the repository

    git clone https://github.com/NusratBegum/Netflix-Recommendation-System-in-Python.git
cd Netflix-Recommendation-System-in-Python

  2. Install required packages

    pip install -r requirements.txt

    Or install packages individually:

    pip install numpy pandas matplotlib seaborn scikit-learn scipy kagglehub jupyter

  3. Configure Kaggle credentials (if needed)

    # Place your kaggle.json in ~/.kaggle/
mkdir -p ~/.kaggle
cp /path/to/kaggle.json ~/.kaggle/
chmod 600 ~/.kaggle/kaggle.json

  4. Launch Jupyter Notebook

    jupyter notebook main.ipynb

Usage

Quick Start

Open main.ipynb and run all cells to:

  1. Download and load the Netflix dataset
  2. Perform comprehensive EDA
  3. Build the recommendation system
  4. Get personalized recommendations

Get Recommendations

# Use the recommendation function
recommend_netflix('Stranger Things')

# Or use the detailed function
recommendations = get_recommendations('Breaking Bad', top_n=10)

Example Output

NETFLIX RECOMMENDATIONS FOR: 'STRANGER THINGS'
================================================

Top 10 Similar Titles:

1. Nightflyers
   Type: TV Show | Genre: TV Horror, TV Mysteries, TV Sci-Fi & Fantasy
   Rating: TV-MA | Year: 2018 | Match: 56.2%

2. Helix
   Type: TV Show | Genre: TV Horror, TV Mysteries, TV Sci-Fi & Fantasy
   Rating: TV-MA | Year: 2015 | Match: 55.3%
...

Project Structure

Netflix-Recommendation-System-in-Python/
│
├── main.ipynb              # Main Jupyter notebook with complete analysis
├── README.md               # Project documentation (this file)
└── .gitignore             # Git ignore file

Notebook Sections

  1. Data Loading & Initial Exploration

    • Library imports
    • Dataset loading
    • Initial data inspection

  2. Feature Types Analysis

    • Data types examination
    • Feature categorization

  3. Data Cleaning & Preprocessing

    • Missing value handling
    • Data type conversions
    • Text preprocessing

  4. Feature Engineering

    • Date feature extraction
    • Duration parsing
    • Genre and country splitting

  5. Exploratory Data Analysis

    • Distribution analysis
    • Temporal trends
    • Geographic insights
    • Genre analysis

  6. Hypothesis Testing

    • Statistical tests
    • Trend validation

  7. Content-Based Recommendation System

    • TF-IDF vectorization
    • Similarity computation
    • Recommendation function

  8. Conclusions & Insights

    • Key findings
    • Summary report

Technologies Used

  • Python 3.8+: Core programming language
  • Pandas: Data manipulation and analysis
  • NumPy: Numerical computing
  • Matplotlib & Seaborn: Data visualization
  • Scikit-learn: Machine learning and NLP
    • TfidfVectorizer: Text feature extraction
    • Cosine Similarity: Content matching
  • SciPy: Statistical analysis
  • Kagglehub: Dataset management
  • Jupyter: Interactive development environment

Key Findings

Content Distribution

  • Movies: 6,131 titles (69.6%)
  • TV Shows: 2,676 titles (30.4%)
  • Netflix has been increasingly adding TV Shows in recent years

Geographic Insights

  • Top Producer: United States (3,211 titles)
  • Other Major Producers: India, United Kingdom
  • Content from 100+ countries represented

Content Characteristics

  • Most Common Rating: TV-MA (Mature Audiences)
  • Top Genres: International Movies, Dramas, Comedies
  • Average Movie Duration: ~100 minutes
  • Average TV Show Seasons: 1.8 seasons

Temporal Patterns

  • Peak content addition: End of year (Q4)
  • Most additions on Fridays
  • Release years span 1925-2021

Recommendation System

How It Works

The recommendation system uses Content-Based Filtering:

  1. Feature Extraction: Combines multiple text features

    • Genres (listed_in)
    • Cast members
    • Director
    • Country
    • Rating
    • Description

  2. Text Vectorization: TF-IDF (Term Frequency-Inverse Document Frequency)

    • 5,000 feature matrix
    • Captures content uniqueness

  3. Similarity Computation: Cosine Similarity

    • Measures content similarity (0-1 scale)
    • Higher scores = more similar content

  4. Recommendation Generation

    • Ranks all titles by similarity
    • Returns top N matches

Performance

  • Matrix Size: 8,807 titles × 5,000 features
  • Computation Time: < 1 second per recommendation
  • Accuracy: High relevance based on content features

Results

The project successfully:

  • Analyzed 8,807 Netflix titles with comprehensive visualizations
  • Identified key trends in Netflix's content strategy
  • Validated hypotheses using statistical tests
  • Built a functional recommendation system
  • Generated relevant recommendations for various content types

Sample Recommendations

For "Stranger Things" (Sci-Fi TV Show):

  • Nightflyers (56.2% match)
  • Helix (55.3% match)
  • Chilling Adventures of Sabrina (54.4% match)

For "Breaking Bad" (Crime Drama):

  • Dare Me (51.1% match)
  • The Lizzie Borden Chronicles (49.9% match)
  • Ozark (47.5% match)

Future Improvements

  • Implement collaborative filtering using user data
  • Add sentiment analysis on descriptions
  • Create hybrid recommendation system
  • Build interactive web application (Flask/Streamlit)
  • Include user ratings and reviews
  • Add real-time trending analysis
  • Implement deep learning models (neural networks)
  • Create API for recommendations

Contributing

Contributions are welcome! Here's how you can help:

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

Nusrat Begum

Acknowledgments

  • Dataset provided by Shivam Bansal on Kaggle
  • Inspired by Netflix's recommendation algorithms
  • Thanks to the open-source community

If you found this project helpful, please consider giving it a star!

Last Updated: December 2025

Languages

Jupyter Notebook100.0%

Contributors

CONU
MIT License
Created December 23, 2025
Updated December 26, 2025