🎥 YouTube Comments Sentiment Analysis

This project uses machine learning to analyze the sentiment of comments on YouTube videos, providing detailed insights and live sentiment overlays via a custom Chrome Extension. Built with modularity and reproducibility in mind, it follows the Cookiecutter Data Science structure and incorporates a full MLOps workflow.

---

📌 Key Features

• ✅ Predicts comment sentiment (Positive / Neutral / Negative)
• 📊 Visual analytics to summarize sentiment trends
• 🌐 Chrome extension that overlays real-time comment sentiment on YouTube
• 🔁 MLOps pipeline with DVC, GitHub Actions, Docker, and Flask API
• 📦 Easy to install, extend, and deploy

---

🧠 Architecture Overview

🔄 Data Pipeline

• YouTube Data API: Collects raw comment data
• Data Preprocessing & Feature Engineering: Cleans and prepares the data for modeling
• Model Training & Evaluation: Sentiment model is trained, validated, and evaluated
• Model Registration: Artifacts are saved and tracked via DVC

🔌 Inference Service

• Flask API (app.py) serves model predictions
• Chrome Extension sends HTTP requests to the API
• Models and vocabulary (vocab.pkl) are versioned and stored in the DVC remote

---

📁 Project Structure

└── youtube-comments-analysis/
    ├── .github
    │   └── workflows
    ├── Dockerfile
    ├── LICENSE
    ├── Makefile
    ├── README.md
    ├── docs
    │   ├── Makefile
    │   ├── commands.rst
    │   ├── conf.py
    │   ├── getting-started.rst
    │   ├── index.rst
    │   └── make.bat
    ├── dvc.lock
    ├── dvc.yaml
    ├── experiment_info.json
    ├── flask_app
    │   ├── app.py
    │   ├── requirements.txt
    │   └── vocab.pkl
    ├── models
    │   └── .gitkeep
    ├── notebooks
    │   ├── .gitkeep
    │   ├── baseline_model.ipynb
    │   ├── baseline_model_bilstm.ipynb
    │   ├── baseline_model_using_lstm_torch.ipynb
    │   ├── bilstm_attention.ipynb
    │   ├── confusion_matrix.png
    │   ├── data_analysis.ipynb
    │   ├── data_ingestion.ipynb
    │   ├── data_preprocessing.ipynb
    │   └── pretrained_models.ipynb
    ├── params.yaml
    ├── references
    │   ├── .gitkeep
    │   └── dataset_resources.txt
    ├── reports
    │   ├── .gitignore
    │   ├── .gitkeep
    │   ├── confusion_matrix.png
    │   └── figures
    ├── requirements.txt
    ├── scripts
    │   └── test_model.py
    ├── setup.py
    ├── src
    │   ├── data
    │   ├── features
    │   ├── model
    │   └── visualization
    ├── test_environment.py
    └── tox.ini

⚙️ Installation & Usage

Step 1: Clone the Repository

git clone https://github.com/Prayesh13/youtube-comments-analysis.git
cd youtube-comments-analysis

Step 2: Set Up Environment

python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

Step 3: Run Pipeline

dvc repro  # run pipeline in powershell

---

🚀 Deploy Inference API

To launch the Flask API for prediction:

python .\flask_app\app.py

The API will be available at http://localhost:5000/predict

Example POST request

POST /predict
{
  "comment": "This video is amazing!"
}

---

🧩 Chrome Extension Integration

This project includes a Chrome extension that overlays sentiment analysis directly onto the YouTube interface.

🔗 Chrome Plugin GitHub Repo: https://github.com/Prayesh13/Yt-chrome-plugin-frontend

Clone the Repository

git clone https://github.com/Prayesh13/Yt-chrome-plugin-frontend.git
cd Yt-chrome-plugin-frontend

📷 Extension Preview

Comment Sentiment UI Comment Sentiment UI

🧪 Testing & CI/CD

• Unit Testing: Configure tests inside tests/ and run with pytest
• CI/CD: GitHub Actions config in .github/workflows/ci-cd.yaml
• Dockerized: Use Dockerfile for containerized deployment

---

📄 License

Distributed under the MIT License. See LICENSE for details.

---

💡 Built with ❤️ using the Cookiecutter Data Science Template