🏷️ LabelMate - AI-Powered Data Labeling Assistant

By @swamy18

A production-ready mini-product that accelerates data annotation workflows for ML teams using AI assistance.

🚀 What LabelMate Does

LabelMate is a comprehensive data labeling tool that combines AI automation with human oversight to create high-quality labeled datasets efficiently:

📄 Text Labeling

• Sentiment Analysis: Automatically classify text as Positive, Negative, or Neutral
• Topic Classification: Categorize text into Technology, Business, Politics, Sports, etc.
• Batch Processing: Label hundreds of texts with one click
• Human Review: Easy interface to correct AI mistakes
• Change Tracking: Monitor which labels were AI-generated vs human-corrected

🖼️ Image Labeling

• Multi-Label Suggestions: AI provides top-3 label suggestions per image
• Flexible Views: Grid, List, or Single-image detailed views
• Custom Labels: Add your own labels beyond AI suggestions
• Batch Operations: Re-process all images with updated AI models

📊 Analytics Dashboard

• Real-time Progress: Track labeling completion rates
• AI Performance: Monitor accuracy and human correction rates
• Data Quality: Insights into label distribution and consistency
• Export Reports: Comprehensive analytics in JSON and Markdown formats

⚙️ Quick Setup (< 5 minutes)

1. Clone and Install

git clone https://github.com/swamy18/labelmate.git
cd labelmate
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Configure API Keys

# Create .env file
echo "GOOGLE_API_KEY=your_gemini_key_here" > .env
# OR for OpenAI
echo "OPENAI_API_KEY=your_openai_key_here" > .env

Get Free API Keys:

• Google Gemini (Recommended - handles both text and images)
• OpenAI (Text only, requires paid account)

3. Launch Application

streamlit run app.py

Open http://localhost:8501 in your browser.

🎯 Demo Workflow

Text Labeling Demo:

1. Upload Data: Use the provided data/sample_texts.csv or upload your own CSV with a 'text' column
2. AI Processing: Click "Auto-Label All" to get AI sentiment predictions
3. Human Review: Review and correct any mislabeled items using the dropdown selectors
4. Export Results: Download the labeled CSV with tracking of AI vs human labels

Image Labeling Demo:

1. Upload Images: Drop 5-10 sample images (JPG/PNG format)
2. AI Suggestions: System provides 3 label suggestions per image using Gemini Vision
3. Label Selection: Choose the best label or add custom ones
4. Export Labels: Download CSV mapping filenames to selected labels

Analytics Demo:

1. Progress Tracking: Real-time charts showing completion rates
2. Performance Metrics: AI accuracy rates and human correction statistics
3. Data Insights: Label distribution analysis and quality metrics
4. Report Export: Generate comprehensive analytics reports

📁 Project Structure

labelmate/
├── app.py                 # Main Streamlit application
├── requirements.txt       # Python dependencies  
├── .env                   # API keys (create this)
├── .gitignore            # Git ignore rules
├── README.md             # This file
├── utils/                # Helper modules
│   ├── __init__.py
│   ├── text_helper.py    # Text labeling logic
│   ├── image_helper.py   # Image labeling logic  
│   └── charts.py         # Analytics and visualization
├── data/                 # Sample data and temporary files
│   ├── sample_texts.csv  # Demo text dataset
│   └── temp_images/      # Temporary image storage
└── exports/              # Output folder (auto-generated)
    ├── labeled_text.csv
    ├── labeled_images.csv
    └── analytics_report.json

🛠️ Technical Stack

• Frontend: Streamlit (Python-based web interface)
• AI Backend: Google Gemini 1.5 Flash (text + vision) or OpenAI GPT-3.5
• Data Processing: Pandas for CSV manipulation
• Visualization: Plotly for interactive charts and progress tracking
• Image Processing: PIL (Pillow) for image handling
• State Management: Streamlit session state for real-time updates

🔧 Advanced Configuration

Multiple AI Providers

The system automatically detects available API keys:

1. Gemini (preferred): Handles both text and image labeling
2. OpenAI: Text labeling only (requires separate vision solution)

Custom Labeling Tasks

Modify utils/text_helper.py to add new classification tasks:

CUSTOM_PROMPT = """
Your custom classification prompt here.
Return one of: Option1, Option2, Option3
"""

Batch Processing Settings

Adjust batch sizes in the UI:

• Text: Process 1-100 items at once
• Images: Upload and process multiple files simultaneously

📈 Production Deployment

Local Production Mode

streamlit run app.py --server.port 8501 --server.address 0.0.0.0

Docker Deployment

FROM python:3.9-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install -r requirements.txt
COPY . .
EXPOSE 8501
CMD ["streamlit", "run", "app.py", "--server.headless", "true"]

Cloud Deployment

• Streamlit Cloud: Direct GitHub integration
• Heroku: Use provided Procfile
• AWS/GCP: Container-based deployment

🤝 Contributing

We welcome contributions! Here's how:

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

Development Setup

# Install development dependencies
pip install -r requirements-dev.txt

# Run tests
python -m pytest tests/

# Code formatting
black app.py utils/

Performance Optimization:

• Text: Process in batches of 10-20 for optimal speed
• Images: Resize large images before upload
• Memory: Clear session state periodically for large datasets

📊 Analytics & Metrics

LabelMate tracks comprehensive metrics:

• Progress Tracking: Real-time completion percentages
• AI Performance: Accuracy rates and confidence scores
• Human Oversight: Correction rates and consistency metrics
• Data Quality: Label distribution and anomaly detection
• Export Analytics: JSON reports for further analysis

🔒 Data Privacy & Security

• Local Processing: All data stays on your machine by default
• API Calls: Only text/image content sent to AI providers (no personal data)
• No Storage: No permanent data storage in cloud services
• Session Based: Data cleared when application restarts

📚 Use Cases

ML/AI Teams

• Dataset Creation: Rapidly label training data for supervised learning
• Data Augmentation: Generate labels for synthetic or scraped data
• Quality Assurance: Validate existing labels with AI assistance

Research Organizations

• Survey Analysis: Classify open-ended survey responses
• Content Analysis: Categorize research papers, articles, or documents
• Image Annotation: Label research images for computer vision projects

Business Applications

• Customer Feedback: Analyze sentiment in reviews and support tickets
• Content Moderation: Classify user-generated content for policy compliance
• Market Research: Tag and categorize social media posts and comments

📄 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

• Streamlit team for the amazing web framework
• Google for the powerful Gemini API
• OpenAI for GPT models and vision capabilities
• Plotly for interactive visualization components
• PIL/Pillow for robust image processing

🌟 Star History

If you find LabelMate useful, please give it a star on GitHub! ⭐

📞 Support & Contact

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: Contact via GitHub profile
• Documentation: Wiki

---

Made by @swamy18

LabelMate - Accelerating AI development through intelligent data labeling """