Maximizing renewable energy utilization for cost savings and environmental impact.
This project addresses the critical problem of sub-optimal renewable energy utilization at Walmart stores. Despite significant investments in renewable energy, challenges persist in maximizing its usage, resulting in missed cost savings and environmental benefits.
Our solution is an ML-powered dashboard that provides actionable insights, enabling Walmart to efficiently manage its diverse energy resources.
Challenge:
Walmart’s large-scale operations generate vast amounts of energy data (consumption, solar generation, grid prices) that require efficient organization and access.
Solution:
- Scalable Data Lake: Built on AWS S3, partitioned for efficient querying.
- Real-Time Data Ingestion: Synthetic data ingested hourly using AWS Lambda triggered by Amazon EventBridge.
- Purpose: Provides a constantly updated dataset for analytics and forecasting.
Tech Stack:
AWS S3, AWS Lambda, Amazon EventBridge
Challenge:
Without predictive foresight, energy decisions are reactive and inefficient.
Solution:
- Forecasting Models:
- Predict future energy consumption and solar generation using time-series forecasting powered by XGBoost.
- Inputs include HVAC, refrigeration, lighting, battery usage, time of day/week, etc.
- Optimization Algorithm:
- Uses forecasted and real-time data to recommend optimal energy mixes (solar, battery, grid).
- Example Recommendation: "Pre-cool the store using solar before peak demand."
Tech Stack:
Python, Pandas, Scikit-learn, XGBoost, NumPy
Challenge:
Need a scalable and efficient method to expose ML-powered intelligence to the dashboard.
Solution:
- FastAPI serves as the backend API:
- Loads ML models for forecasts and optimization.
- Provides endpoints for real-time metrics, forecasts, recommendations, and historical data querying.
- Deployment: Hosted on Railway.com using Docker containers for consistency.
Tech Stack:
FastAPI, Docker, boto3 (for AWS S3 interaction)
Challenge:
Store managers need clear, actionable insights presented visually for informed decisions.
Solution:
- Frontend Dashboard:
- Built with Next.js and TypeScript.
- Features real-time energy consumption breakdown, interactive forecasts, optimization recommendations, and historical data.
- Fully responsive and interactive UI.
Tech Stack:
Next.js, TypeScript, React, Tailwind CSS, Recharts
| Component | Technology | Hosting |
|---|---|---|
| Data Lake | AWS S3 | AWS |
| Data Pipeline | AWS Lambda + EventBridge | AWS |
| Backend API + ML Model | FastAPI + Docker | Railway |
| Frontend UI | Next.js + Tailwind | Vercel |
- ✅ Real-time renewable energy utilization display
- ✅ Predictive energy forecasting
- ✅ Intelligent optimization recommendations
- ✅ Visualization of historical energy data
- ✅ Fully responsive, interactive web interface
- Integration with real-time IoT sensor data from actual Walmart stores
- Advanced optimization algorithms (e.g., Reinforcement Learning)
- Cost savings estimation module
- Authentication & Role-based access control (RBAC)
Contributions are welcome! Please fork the repository and submit a pull request.
- Ratnesh Kumar Jaiswal
- Contact: ratnesh.kr.jais@gmail.com | LinkedIn