Skip to content

README.md

Latest commit

 

History

History
94 lines (68 loc) · 2.34 KB

README.md

File metadata and controls

94 lines (68 loc) · 2.34 KB

Sentiment Analyzer for Product Reviews

A full-stack web application that analyzes the sentiment of user-submitted reviews or comments (e.g., from Amazon or Twitter) and classifies them as Positive, Neutral, or Negative. Built using Python (Flask) for the backend and React for the frontend.

Features

  • Train and save a logistic regression model using scikit-learn
  • REST API built with Flask to serve real-time predictions
  • React frontend for user input and displaying sentiment results
  • CORS-enabled communication between frontend and backend

Tech Stack

  • Frontend: React, Axios
  • Backend: Python, Flask, scikit-learn, joblib, Flask-CORS
  • Model: Logistic Regression with TF-IDF vectorizer
  • Tools: Postman (for API testing), VS Code (for dev)

Project Structure

sentiment-analyzer/
├── backend/
│   ├── app.py                  # Flask app with /predict API
│   ├── model/
│   │   ├── train_model.py      # Model training script
│   │   ├── sentiment_model.pkl # Saved ML model
│   │   └── sentiment_data.csv  # CSV with labeled text data
│   └── requirements.txt        # Python dependencies
├── frontend/
│   └── sentiment-ui/           # React app
└── README.md

Setup Instructions

1. Backend (Flask API)

cd backend
python -m venv venv
source venv/bin/activate      # On Windows: venv\Scripts\activate
pip install -r requirements.txt
cd model
python train_model.py         # Trains and saves model
cd ..
python app.py                 # Starts Flask server at http://127.0.0.1:5000

2. Frontend (React App)

cd frontend
npx create-react-app sentiment-ui
cd sentiment-ui
npm install axios
npm start                     # Starts frontend at http://localhost:3000

Example Usage

  1. Open the app at http://localhost:3000
  2. Enter a review like:
    "This is the best product I’ve ever used!"
  3. Click "Analyze Sentiment"
  4. Output:
    Sentiment: POSITIVE

Future Enhancements

  • Upgrade to Hugging Face Transformers (DistilBERT) for better accuracy
  • Add charts for sentiment analysis stats
  • Enable batch predictions
  • Dockerize and deploy to Render or Railway
  • Add CI/CD with GitHub Actions

License

This project is open-source and free to use under the MIT license.