Campus Energy forecast- ML powered dashboard for hostel energy optimization
This project aims to develop a web-based dashboard that uses hybrid machine learning models to predict short-term energy consumption, detect anomalies, and optimize electricity costs specifically for university hostels and college buildings.
WattWise – AI-Powered Smart Campus Energy Forecaster
### Problem Statement
Indian university hostels and college buildings are facing a severe energy crisis. Unpredictable usage patterns — driven by heavy air conditioning during scorching summers, overnight fans and lights, and high-consumption mess/kitchen operations — result in skyrocketing electricity bills. Traditional manual monitoring methods are reactive, inefficient, and often inaccurate. This leads to 20–40% avoidable energy wastage, frequent overload penalties, inflated costs, and significant challenges in achieving sustainability goals on campus. Without intelligent insights, administrators struggle to identify inefficiencies, predict demand, or implement timely corrective actions.
### The Solution: WattWise
WattWise is a modern, full-stack AI-powered energy intelligence dashboard specifically designed for educational institutions in India. It delivers real-time monitoring, accurate short-term forecasting using hybrid machine learning models, automatic anomaly detection, and smart billing simulations — all presented in an engaging, cyberpunk-themed, responsive web interface.
Developed as a FossHack project with clean, expert-level, and production-ready code quality, WattWise empowers campus administrators, wardens, and facility managers to shift from reactive firefighting to proactive, data-driven energy management. The platform helps reduce electricity bills by an estimated 15–30% through early warnings, behavioral nudges, and optimization insights while supporting green campus initiatives.
### Core Features (MVP – All 5 Completed)
1. Secure User Authentication & Role Management
Robust JWT-based registration and login system with clear role-based access (Admin and Viewer). Admins get a dedicated user management panel to promote, demote, or delete users, ensuring secure and controlled access.
2. Seamless CSV Data Upload & Storage
An intuitive drag-and-drop interface allows admins to upload historical meter and sub-meter readings. The system supports structured data including date, block/building, room, appliance, and kWh consumption. Data is stored efficiently (currently JSON-based, easily upgradeable to SQLite).
3. Real-Time & Historical Usage Monitoring
Live tracking of energy consumption across different campus blocks and hostels. Interactive 7-day trend charts built with Chart.js provide clear visualizations of usage patterns over time.
4. Advanced Short-Term Predictive Forecasting
A hybrid LSTM + XGBoost machine learning model delivers accurate next-day, block-wise energy predictions along with confidence scores. A dedicated LSTM Forecast Dashboard displays model performance metrics such as MAE and RMSE, helping users trust and interpret forecasts effectively.
5. Intelligent Anomaly Detection & Alerts
The system automatically identifies unusual consumption patterns (for example, excessive usage at odd hours) and displays visual alerts with different severity levels for quick action.
### Additional Standout Features
- Gemini-Powered AI Energy Assistant: A multi-turn conversational chatbot that answers queries, provides recommendations, and explains insights in natural language.
- Billing & Budget Management: Real-time cost calculations (based on user-input ₹/kWh rates), monthly bill projections, visual budget progress bars, and “what-if” simulation tools.
- Detailed Block-wise Analysis: Separate dashboards for Girls Hostel, Boys Hostel, Academic Blocks (AB1, AB2), and Admin Building with deep consumption insights.
- Report Export: One-click CSV export for reports, forecasts, and billing data.
- Modern Cyberpunk UI: Fully responsive dark-themed design with smooth animations, offering an excellent user experience on both desktop and mobile devices.
### Tech Stack
- Frontend: HTML5, CSS3 (custom cyberpunk design system), Vanilla JavaScript, Chart.js v4
- Backend: Node.js + Express.js, JWT authentication, bcryptjs for password security
- AI/ML Integration: Secure Gemini API proxy for the AI assistant; ready endpoints for hybrid LSTM + XGBoost forecasting
- Development Tools: Nodemon, dotenv, CORS support
### Impact & Hackathon Relevance
WattWise directly addresses real-world pain points faced by Indian campuses. By combining real-time visibility, predictive analytics, and AI assistance, it promotes energy efficiency, reduces wastage, lowers operational costs, and contributes to sustainability targets. The project demonstrates end-to-end development skills — from secure backend and intuitive frontend to machine learning integration and polished UI/UX.
Built with clean architecture and production-ready standards in a short timeframe, WattWise is fully open-source, well-documented, and extensible for future enhancements such as IoT sensor integration, advanced Python ML pipelines, or SQLite database migration.
WattWise stands as a practical, impactful solution that can make university campuses smarter, greener, and more cost-efficient.