Biz-Pulse

BizPulse is an AI-powered business intelligence co-pilot for SMBs. It turns raw sales, feedback, and market data into clear narrative insights, anomaly alerts, and action-oriented recommendations — helping small businesses make faster, smarter, data-driven decisions.

AWS Integration Architecture

BizPulse leverages multiple AWS services to provide a comprehensive AI-powered business intelligence platform. Here's how AWS is integrated throughout the application:

1. Amazon S3 - Data Lake Storage

• Primary Use: Centralized data storage for all business data files
• Implementation:
  • Configured as the default file storage backend in Django settings
  • Stores uploaded CSV, Excel, and JSON files in organized folder structures (raw-uploads/)
  • Environment variables: AWS_STORAGE_BUCKET_NAME, AWS_S3_REGION_NAME
  • Custom domain setup: https://{bucket-name}.s3.amazonaws.com/
• Features:
  • File metadata storage (job IDs, timestamps)
  • Direct integration with Django's file handling system
  • CORS-enabled for frontend access

2. AWS Glue - ETL Processing

• Primary Use: Serverless data processing and transformation
• Implementation:
  • Custom utility functions in core/aws_utils.py for job orchestration
  • Asynchronous job execution with status polling
  • Integration with S3 for input/output data
• Features:
  • Automated data cleansing and transformation
  • Real-time job status monitoring
  • Error handling and retry mechanisms

3. Amazon Bedrock - AI-Powered Insights

• Primary Use: Natural language processing for business intelligence
• Implementation:
  • Uses Anthropic Claude 3 Haiku model via Bedrock Runtime API
  • Located in services/analysis.py
  • Processes structured business metrics to generate narrative insights
• Features:
  • Automated insight generation from KPI data
  • JSON-structured output with titles, summaries, and recommendations
  • Context-aware analysis based on data source metadata
  • Integration with Django ORM for insight storage

4. Amazon Q Business - Natural Language Queries

• Primary Use: Conversational AI for business questions
• Implementation:
  • Anonymous access application in services/amazon_q_service.py
  • Environment variables: AMAZON_Q_APP_ID, AMAZON_Q_REGION
  • REST API endpoint for natural language queries
• Features:
  • Source attribution for answers
  • Conversation context tracking
  • Error handling for failed queries

5. Amazon Forecast - Predictive Analytics

• Primary Use: Time series forecasting for business metrics
• Implementation:
  • Internal API endpoints in internal_api/views.py
  • Test framework in testForecast.py
  • Stores predictions in structured database models
• Features:
  • Automated forecast generation
  • Integration with insight system for narrative summaries
  • Historical data analysis for trend prediction

6. Amazon Lookout for Metrics - Anomaly Detection

• Primary Use: Real-time anomaly detection in business metrics
• Implementation:
  • Internal API endpoints for anomaly ingestion
  • Test framework in anomalyTest.py
  • Severity scoring and automated alerting
• Features:
  • Real-time monitoring of KPI deviations
  • Automated insight generation for detected anomalies
  • Integration with alerting system

7. AWS Lambda - Serverless Computing

• Primary Use: Event-driven processing and microservices
• Implementation:
  • Serverless functions for data processing workflows
  • Integration with S3 events and API Gateway
  • Asynchronous task execution for heavy computations
• Features:
  • Auto-scaling based on workload
  • Cost-effective compute for sporadic tasks
  • Integration with other AWS services via event triggers

AWS Configuration

Environment Variables Required:

# S3 Configuration
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key
AWS_STORAGE_BUCKET_NAME=bizpulse-data-lake
AWS_S3_REGION_NAME=us-east-1

# Amazon Q Business
AMAZON_Q_APP_ID=your_app_id
AMAZON_Q_REGION=us-east-1

# Bedrock
AWS_REGION_NAME=us-east-1

# Database (optional)
DATABASE_URL=your_database_url

Django Settings Integration:

• Storage Backend: S3Boto3Storage configured for all file operations
• CORS: Enabled for cross-origin requests from frontend
• Authentication: JWT-based authentication with AWS-integrated services

Data Flow Architecture

1. Data Ingestion: Files uploaded via frontend → S3 storage → Glue processing
2. AI Analysis: Processed data → Bedrock for insights → Database storage
3. Query Processing: Natural language queries → Amazon Q → Response with sources
4. Predictive Analytics: Historical data → Forecast models → Prediction storage
5. Anomaly Detection: Real-time metrics → Lookout for Metrics → Alert generation

Benefits of AWS Integration

• Scalability: Serverless architecture handles variable workloads
• Cost Efficiency: Pay-per-use model for AI and storage services
• Security: Enterprise-grade security with AWS IAM and encryption
• Reliability: Multi-AZ deployment and automatic failover
• Innovation: Access to cutting-edge AI models and analytics tools

This comprehensive AWS integration enables BizPulse to deliver enterprise-grade business intelligence capabilities to small and medium businesses at a fraction of the traditional cost.

Backend Architecture

The backend is built with Django REST Framework and deployed on Render, providing a robust API-first architecture for the business intelligence platform.

Key Technologies:

• Django 4.2 - Web framework with REST API capabilities
• Django REST Framework - API development toolkit
• PostgreSQL - Primary database for user data, insights, and analytics
• JWT Authentication - Secure token-based authentication
• AWS Integration - Multiple AWS services for AI and data processing

Deployment:

• Platform: Render (render.com)
• Database: PostgreSQL hosted on Render
• Environment: Production-ready with automatic scaling
• API Documentation: OpenAPI/Swagger integration via DRF Spectacular

Core Components:

• User Management: Registration, authentication, and profile management
• Data Services: File upload, processing, and analytics APIs
• AI Integration: Amazon Bedrock, Q Business, and predictive analytics
• Internal APIs: Forecast and anomaly detection endpoints

Frontend Architecture

The frontend is a modern React application deployed on Vercel, providing an intuitive user interface for business intelligence dashboards and data management.

Key Technologies:

• React 18 - Component-based UI framework
• Vite - Fast build tool and development server
• Tailwind CSS - Utility-first CSS framework
• React Router - Client-side routing
• Axios - HTTP client for API communication
• JWT Authentication - Secure frontend authentication flow

Deployment:

• Platform: Vercel (vercel.com)
• Build Process: Automated deployments from Git
• Environment: Global CDN with edge computing
• Domain: Custom domain support

Core Features:

• Dashboard: Real-time business metrics and insights
• Data Connections: File upload and integration management
• Sales Analytics: Interactive charts and deep-dive analysis
• User Authentication: Login/register with JWT token management
• Responsive Design: Mobile-first approach with Tailwind CSS

Technology Stack Summary

Component	Technology	Deployment	Purpose
Backend	Django REST Framework	Render	API services, AI integration
Frontend	React + Vite	Vercel	User interface, dashboards
Database	PostgreSQL	Neon	Data persistence
File Storage	Amazon S3	AWS	Data lake, file uploads
AI/ML	Amazon Bedrock, Q Business	AWS	Insights, natural language
ETL	AWS Glue	AWS	Data processing
Analytics	Amazon Forecast, Lookout	AWS	Predictions, anomaly detection