🛒 RetailIQ - AI Copilot for Demand Forecasting

AI-powered demand forecasting and inventory optimization platform for small retailers in Bharat.

Features

• 📤 CSV Data Upload - Easy sales data ingestion
• 📈 Demand Forecasting - 7/14/30 day predictions with Prophet ML
• 📦 Inventory Optimization - Automated reorder recommendations
• 🎯 Stockout Risk Scoring - Prioritized alerts (0-100 scale)
• 🤖 AI Copilot - Natural language Q&A with Amazon Bedrock
• 🇮🇳 Bharat-Specific - Festival dates, regional factors, ₹ formatting

Tech Stack

• Backend: Python, FastAPI, boto3
• ML: Prophet, scikit-learn, pandas, numpy
• AI: Amazon Bedrock (Claude 3 Sonnet)
• AWS: Lambda, S3, DynamoDB, API Gateway
• Frontend: HTML/CSS/JavaScript, Bootstrap, Chart.js

Quick Start

---

1. Install Dependencies

pip install -r requirements.txt

---

2. Configure AWS

aws configure

1. Enter your AWS credentials
2. Region: ap-south-1

---

3. Setup AWS Resources

python config/aws_setup.py

---

4. Run API Server

python -m api.main

---

5. Run Frontend

• Open the terminal:

cd frontend python -m http.server 3000

---

6. Open Browser

http://localhost:3000

---

Project Structure

retailiq-prototype/ ├── api/ # FastAPI application ├── lambda_functions/ # AWS Lambda handlers │ ├── data_ingestion/ │ ├── forecasting/ │ ├── inventory/ │ └── copilot/ ├── utils/ # Shared utilities │ ├── csv_validator.py │ ├── forecasting.py │ ├── inventory_optimizer.py │ ├── regional_factors.py │ └── copilot.py ├── config/ # AWS setup scripts ├── frontend/ # Web dashboard ├── data/ # Sample data └── tests/ # Test files

API Endpoints

POST /data/upload - Upload sales CSV POST /forecast - Generate demand forecast POST /inventory/recommendations - Get reorder recommendations POST /copilot/chat - AI Copilot Q&A GET /health - Health check

Demo Credentials

User ID: demo_user API Key: test-key-123

AWS Resources Created

S3 Buckets:

• retailiq-raw-data-varun
• retailiq-processed-data-varun
• retailiq-model-artifacts-varun

DynamoDB Tables:

• retailiq-metadata
• retailiq-model-registry

Cost Estimate

Free Tier: Lambda, S3, DynamoDB, API Gateway (covered) Bedrock: ~$0.03-0.10 for hackathon testing Total: < $1 for 3-day hackathon

Hackathon Submission

Team: Coding Gladiators Event: AI for Bharat Hackathon
Category: Retail AI Solutions

Key Highlights

• Solves real problem for Bharat retailers
• Uses AWS AI services (Bedrock)
• Serverless architecture (scalable, cost-effective)
• Regional customization (festivals, languages)
• Production-ready architecture
• Responsible AI practices

Future Enhancements

• Mobile app (React Native)
• WhatsApp alerts for retailers
• Multi-language UI (Hindi, Tamil, Telugu)
• Advanced ML models (XGBoost, LSTM)
• POS system integration
• Pricing optimization recommendations