Bluesky Sentiment Analysis

A full-stack web application that analyzes sentiment of posts (skeets) on Bluesky in real-time. The platform uses advanced AI models to classify public sentiment on any topic of interest.

Overview

Bluesk is a sentiment analysis tool built for the Bluesky social network. It aggregates posts about a specific topic and uses machine learning to determine whether the overall sentiment is positive, neutral, or negative. Perfect for tracking public opinion, monitoring brand sentiment, or analyzing trending topics.

Features

• 🔍 Real-time Post Collection - Fetch relevant skeets from Bluesky using search queries
• 🤖 AI-Powered Sentiment Analysis - Classify posts using cardiffnlp/twitter-roberta-base-sentiment model
• 📊 Visual Analytics - Interactive charts and visualizations of sentiment distribution
• 📥 Data Export - Download analysis results in multiple formats
• 🎨 Modern UI - Beautiful, responsive interface built with Next.js and Tailwind CSS
• ⚡ Fast Performance - Optimized backend with FastAPI and efficient text processing

Tech Stack

Frontend

• Framework: Next.js 15 with React 19
• Styling: Tailwind CSS with custom animations
• Components: Radix UI component library
• Visualization: Recharts
• Animation: Framer Motion
• Form Handling: React Hook Form with Zod validation
• UI Library: shadcn/ui

Backend

• API: FastAPI (Python)
• Bluesky Integration: atproto library
• NLP: Hugging Face Transformers (cardiffnlp/twitter-roberta-base-sentiment)
• CORS: Enabled for frontend communication

Project Structure

bluesk/
├── app/                      # Next.js app directory
│   ├── page.tsx             # Home/landing page
│   ├── layout.tsx           # Root layout
│   ├── globals.css          # Global styles
│   ├── search/              # Search page
│   └── results/             # Results page
├── components/              # Reusable React components
├── lib/                     # Utility functions
├── hooks/                   # Custom React hooks
├── main.py                  # FastAPI backend server
├── package.json             # Node dependencies
├── tailwind.config.js       # Tailwind configuration
├── tsconfig.json            # TypeScript configuration
└── next.config.mjs          # Next.js configuration

Installation

Prerequisites

• Node.js 18+
• Python 3.9+
• Bluesky account credentials

Frontend Setup

npm install
# or
npm install --legacy-peer-deps

Backend Setup

pip install fastapi transformers atproto python-dotenv

Configuration

Create a .env file in the root directory:

BSKY_USERNAME=your_bluesky_username
BSKY_PASSWORD=your_bluesky_password

Usage

Development

Start the frontend (Next.js):

npm run dev

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

Start the backend (FastAPI):

python main.py

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

Production Build

npm run build
npm start

How It Works

1. Search - User enters a topic of interest on the search page
2. Collection - The backend queries Bluesky's API for up to 100 posts matching the search term
3. Analysis - Each post is processed and analyzed using the sentiment analysis model
   • Text is cleaned (URLs, mentions, special characters removed)
   • Sentiment scores are calculated (positive, neutral, negative)
   • Dominant sentiment is determined
4. Visualization - Results are displayed with:
   • Sentiment distribution charts
   • Individual post sentiment tags
   • Engagement metrics (likes, reposts)
   • Sortable/filterable post list

API Endpoints

/api/analyze

Analyzes posts for a given search term.

Query Parameters:

• term (string, required) - The search term to analyze

Response:

{
  "data": [
    {
      "id": "post_uri",
      "text": "post content",
      "sentiment": {
        "sentiment": "positive|neutral|negative",
        "score": 0.95
      },
      "created_at": "2024-01-01T12:00:00Z",
      "likes": 42,
      "retweets": 10
    }
  ]
}

License

MIT License - see LICENSE file for details

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Future Enhancements

• Advanced filtering options (date range, engagement filters)
• Historical sentiment tracking
• Custom model training
• Multi-language support
• Real-time sentiment streaming
• User authentication and saved searches