Twitter Sentiment Analysis

A full-stack application for analyzing sentiment in tweets and text using AI models. Developed by Hüseyin ASLIM as part of the Başkent University BIL479 - Pattern Recognition course project.

Features

• Text Sentiment Analysis
• Tweet URL Sentiment Analysis
• Twitter User Sentiment Analysis
• Visualization of sentiment results
• Multi-language support (English, Turkish, and more)
• Modern UI with light and dark mode

Model Information

The project uses the cardiffnlp/twitter-xlm-roberta-base-sentiment model from the Hugging Face Transformers library. This model:

• Is based on XLM-RoBERTa architecture
• Supports multilingual sentiment analysis
• Classifies text into three categories: Positive, Neutral, and Negative
• Provides confidence scores for each sentiment category

Tech Stack

Backend

• FastAPI
• Transformers
• PyTorch
• Pandas
• Matplotlib
• Playwright (for web scraping)
• ntscraper (for Twitter scraping)

Frontend

• Next.js
• React
• Tailwind CSS
• shadcn/ui (UI components)
• Axios (API client)

Project Structure

├── backend/             # FastAPI backend
│   ├── main.py          # Main API file
│   └── requirements.txt # Python dependencies
│
└── frontend/            # Next.js frontend
    ├── public/          # Static assets
    ├── src/             # Source code
    │   ├── app/         # Next.js app router
    │   ├── components/  # React components
    │   ├── lib/         # Utilities
    │   └── types/       # TypeScript types
    └── package.json     # Node dependencies

Setup Instructions

Backend

1. Navigate to the backend directory:

   cd backend

2. Install dependencies:

   pip install -r requirements.txt

3. Install Playwright browsers:

   playwright install

4. Start the server:

   uvicorn main:app --reload

The API will be available at http://localhost:8000

Frontend

1. Navigate to the frontend directory:

   cd frontend

2. Install dependencies:

   npm install

3. Start the development server:

   npm run dev

The frontend will be available at http://localhost:3000

API Endpoints

Health Check

• GET /: Check if the API is running

Text Analysis

• POST /analyze/text: Analyze sentiment of provided text
  • Body: {"text": "Your text here"}
  • Returns: Sentiment label, confidence scores, and visualization

Tweet Analysis

• POST /analyze/tweet: Analyze sentiment of a tweet from URL
  • Body: {"url": "https://twitter.com/username/status/tweetid"}
  • Returns: Tweet content, sentiment label, confidence scores, and visualization

User Analysis

• POST /analyze/user: Analyze sentiment of tweets from a specific user
  • Body: {"username": "username", "max_results": 20}
  • Returns: User statistics, sentiment distribution, top positive/negative tweets with charts

Deployment

Backend Deployment

The backend can be deployed to any server that supports Python:

1. Clone the repository
2. Install the requirements: pip install -r backend/requirements.txt
3. Install Playwright browsers: playwright install
4. Run with Gunicorn (for production):

   gunicorn -w 4 -k uvicorn.workers.UvicornWorker backend.main:app

Frontend Deployment

The Next.js frontend can be deployed to Vercel or any hosting provider:

1. Build the application:

   cd frontend
   npm run build

2. Start the production server:

   npm start

Alternatively, deploy directly to Vercel:

vercel

Requirements

Backend Requirements

• Python 3.9+
• FastAPI 0.104.1+
• Transformers 4.35.0+
• PyTorch
• Playwright 1.39.0+
• Other dependencies listed in backend/requirements.txt

Frontend Requirements

• Node.js 18+
• Next.js 14+
• React 18+
• Other dependencies listed in frontend/package.json