Next-Generation AI-Powered Sales Call Analysis System

Transform your sales conversations into actionable insights with advanced ML, real-time sentiment analysis, and intelligent lead scoring. Built with modern architecture for enterprise scalability and cloud deployment.

• Multi-Model Sentiment Analysis - Ensemble approach with VADER, TextBlob, and Transformers
• Intelligent Lead Scoring - AI-powered classification with confidence metrics
• Topic Extraction - Automatic identification of conversation themes
• Interest Detection - High-engagement moment identification
• Speaker Profiling - Role identification and engagement analysis

• Real-time Analytics - Live processing status and performance metrics
• Interactive Visualizations - Responsive charts and data exploration
• Advanced Filtering - Multi-criteria search and sorting capabilities
• Export Capabilities - PDF reports and structured data export
• Mobile Responsive - Optimized for all device types

• Cloud-Native Design - AWS, Azure, GCP compatible
• Microservices Architecture - Scalable and maintainable components
• Container-Ready - Docker and Kubernetes deployment support
• API-First Design - RESTful APIs with comprehensive documentation
• Security Hardened - Enterprise-grade security practices

System Requirements:

• Python 3.11+
• Node.js 18.0+
• Docker & Docker Compose (recommended)
• Git

Required Services:

• AWS Account (for S3 and optional Transcribe)
• OpenAI API Key (optional, for enhanced analysis)

# Clone the repository
git clone https://github.com/aaron-seq/ML-voice-lead-analysis.git
cd ML-voice-lead-analysis

# Configure environment
cp backend/.env.example backend/.env
cp frontend/.env.example frontend/.env.local

# Edit configuration files with your settings
# nano backend/.env
# nano frontend/.env.local

# Start all services
docker-compose up --build

# Access the application:
# Frontend: http://localhost:3000
# Backend API: http://localhost:8000
# API Documentation: http://localhost:8000/v1/docs

# Backend setup
cd backend
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt
python -m spacy download en_core_web_md

# Frontend setup (new terminal)
cd frontend
npm install

# Start services
# Terminal 1 - Backend
cd backend
uvicorn main:app --reload --host 0.0.0.0 --port 8000

# Terminal 2 - Frontend  
cd frontend
npm run dev

# Application Settings
ENVIRONMENT=development
DEBUG=true
PORT=8000

# AWS Configuration
AWS_REGION=us-east-1
DATA_BUCKET=your-s3-bucket-name
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key

# Database (choose one)
DATABASE_URL=sqlite+aiosqlite:///./voice_analysis.db
# DATABASE_URL=postgresql+asyncpg://user:pass@localhost:5432/voice_analysis

# Security
SECRET_KEY=your-super-secret-key-minimum-32-characters

# ML Configuration
USE_TRANSFORMERS=false  # Set to true for enhanced analysis
CLOUD_DEPLOYMENT=false  # Set to true for cloud optimization

# API Configuration
NEXT_PUBLIC_API_URL=http://localhost:8000
NEXT_PUBLIC_APP_NAME=ML Voice Lead Analysis
NEXT_PUBLIC_VERSION=3.1.0

# Environment
NODE_ENV=development
NEXT_TELEMETRY_DISABLED=1

graph TB
    A[Frontend Dashboard<br/>React/Next.js] --> B[API Gateway<br/>FastAPI]
    B --> C[Analysis Service<br/>ML Pipeline]
    B --> D[AWS S3<br/>File Storage]
    B --> E[Database<br/>PostgreSQL/SQLite]
    B --> F[Cache<br/>Redis]
    C --> G[NLP Models<br/>spaCy, Transformers]
    C --> H[Sentiment Analysis<br/>Multi-model Ensemble]
    C --> I[Lead Scoring<br/>AI Classification]

• Frontend: Modern React with TypeScript, Tailwind CSS, and Next.js
• Backend: FastAPI with async processing, SQLAlchemy ORM, structured logging
• ML Pipeline: Advanced NLP with spaCy, TensorFlow, Transformers
• Storage: AWS S3 for files, PostgreSQL for data, Redis for caching
• Infrastructure: Docker containers, cloud deployment ready

1. Upload Audio Transcript to your S3 bucket:

   s3://your-bucket/transcripts/sample-call.json
   

2. Run Analysis Pipeline:

   cd pipeline
   python voice_analysis_pipeline.py transcripts/sample-call.json

3. View Results in the dashboard at http://localhost:3000

import requests

# Retrieve paginated call list
response = requests.get(
    "http://localhost:8000/v1/calls",
    params={"page": 1, "page_size": 10, "sort_field": "upload_timestamp"}
)
calls = response.json()

print(f"Found {calls['pagination']['total_items']} calls")

# Get comprehensive analysis
response = requests.get("http://localhost:8000/v1/calls/sample-call.json")
analysis = response.json()

print(f"Lead Score: {analysis['lead_score_details']['primary_classification']}")
print(f"Sentiment: {analysis['sentiment_score']:.2f}")
print(f"Key Topics: {', '.join(analysis['extracted_topics'])}")

# Reprocess with latest models
response = requests.post("http://localhost:8000/v1/calls/sample-call.json/reprocess")
result = response.json()

print(f"Status: {result['status']}")
print(f"Processing ID: {result['processing_id']}")

# Install Vercel CLI
npm i -g vercel

# Configure environment variables in Vercel dashboard:
# - AWS_ACCESS_KEY_ID
# - AWS_SECRET_ACCESS_KEY
# - S3_BUCKET_NAME
# - SECRET_KEY

# Deploy
vercel --prod

1. Backend Service:

   • Connect your GitHub repository
   • Set build command: pip install -r backend/requirements.txt
   • Set start command: cd backend && uvicorn main:app --host 0.0.0.0 --port $PORT
   • Add environment variables

2. Frontend Service:

   • Set build command: cd frontend && npm install && npm run build
   • Set start command: cd frontend && npm start
   • Add environment variables

# Install Railway CLI
npm install -g @railway/cli

# Login and deploy
railway login
railway link
railway up

# Package backend for Lambda
cd backend
pip install -r requirements.txt -t lambda_package/
cp -r . lambda_package/
cd lambda_package && zip -r ../lambda_deployment.zip .

# Deploy using AWS CLI or Console
aws lambda update-function-code \
  --function-name ml-voice-analysis \
  --zip-file fileb://lambda_deployment.zip

# Build production images
docker-compose -f docker-compose.yml build

# Deploy with production settings
ENVIRONMENT=production \
SECRET_KEY=your-production-secret \
AWS_ACCESS_KEY_ID=your-key \
AWS_SECRET_ACCESS_KEY=your-secret \
docker-compose up -d

# Backend tests
cd backend
pytest tests/ -v --coverage

# Frontend tests
cd frontend
npm test

# Integration tests
docker-compose -f docker-compose.test.yml up --abort-on-container-exit

# Backend formatting and linting
cd backend
black .
isort .
flake8 .
mypy .

# Frontend formatting and linting
cd frontend
npm run lint
npm run format
npm run type-check

Monitor application performance through:

• Health Check: GET /health
• Performance Metrics: GET /v1/analytics/performance
• System Info: GET /v1/system/info

• JWT-based authentication support
• Input validation and sanitization
• SQL injection prevention
• CORS configuration
• Rate limiting ready
• Container security best practices

# Strong secret key (32+ characters)
SECRET_KEY=your-cryptographically-secure-secret-key

# Database connection with SSL
DATABASE_URL=postgresql+asyncpg://user:pass@host:5432/db?ssl=require

# CORS origins (production)
CORS_ORIGINS=https://yourdomain.com,https://api.yourdomain.com

• Async/await for all I/O operations
• Connection pooling for databases
• Multi-layer caching with Redis
• Request batching and pagination
• Background task processing

• Code splitting and lazy loading
• Image optimization with Next.js
• Bundle analysis and optimization
• Performance monitoring
• CDN-ready static assets

• Model caching and reuse
• Batch processing capabilities
• Memory-efficient processing
• GPU acceleration support (optional)
• Lightweight models for cloud deployment

# Check Python version
python --version  # Should be 3.11+

# Verify dependencies
pip install -r requirements.txt

# Download spaCy model
python -m spacy download en_core_web_md

# Check environment variables
cat backend/.env

# Clear cache and reinstall
npm cache clean --force
rm -rf node_modules package-lock.json
npm install

# Check Node version
node --version  # Should be 18+

# Verify AWS credentials
aws sts get-caller-identity

# Test S3 access
aws s3 ls s3://your-bucket-name

# Check bucket permissions
aws s3api get-bucket-policy --bucket your-bucket-name

• Slow API Response: Check database query performance, Redis cache hit rates
• High Memory Usage: Adjust ML model batch sizes, implement model unloading
• Container Issues: Monitor resource limits, check container logs

• 📖 Documentation: Comprehensive guides and API docs
• 🐛 Issues: GitHub Issues
• 💬 Discussions: GitHub Discussions
• 📧 Contact: aaronsequeira12@gmail.com

1. Fork the repository
2. Create feature branch: git checkout -b feature/amazing-feature
3. Make changes with comprehensive tests
4. Run test suite: npm test and pytest
5. Commit changes: git commit -m 'feat: add amazing feature'
6. Push to branch: git push origin feature/amazing-feature
7. Open Pull Request

• Python: Black formatting, isort imports, flake8 linting, type hints
• TypeScript: Prettier formatting, ESLint, strict type checking
• Commit Messages: Conventional Commits format
• Documentation: Comprehensive docstrings and README updates

This project is licensed under the MIT License - see the LICENSE file for details.

• spaCy Team for excellent NLP libraries
• FastAPI for the modern Python web framework
• Next.js Team for the powerful React framework
• Vercel for seamless deployment platform
• Open Source Community for incredible tools and libraries

• Lines of Code: 20,000+
• Test Coverage: 90%+
• API Endpoints: 25+
• ML Models: 5+ integrated
• Deployment Platforms: 8+ supported
• Processing Speed: ~2-3 minutes per call
• Accuracy: 85%+ lead scoring accuracy

Built with ❤️ by Aaron Sequeira

⭐ Star this repo | 🐛 Report Bug | ✨ Request Feature

Transform your sales conversations into actionable insights with enterprise-grade AI analysis.