the front - page where you upload your bill
the problem and wastage of the power/electricity
if we fix how much you save electricity, money
the proof of it and down below you can even get the report car printed

WattWise — Energy Waste Detective

Tagline: Turn electricity data into actionable energy savings.

Inspiration

Electricity bills show how much energy we consume, but they rarely explain when and where energy may be getting wasted. Unusual overnight consumption, unexpected usage spikes, and inefficient energy patterns can go unnoticed.

We built WattWise to help users understand their electricity consumption and identify opportunities to save energy using data-driven analysis.

What it does

WattWise is an energy-analysis application that transforms electricity consumption data into understandable insights. Users can upload energy readings, analyze consumption patterns, investigate unusual usage, and explore potential cost savings.

The application combines a Python-based calculation engine with an interactive user interface. Where the AI integration is enabled, an NVIDIA Nemotron model helps translate analytical results into clear explanations and practical recommendations.

Key features

Energy Data Analysis: Upload CSV files containing timestamps and electricity consumption readings.
Consumption Insights: Calculate total energy usage and estimated electricity costs.
Overnight Baseline Analysis: Compare overnight consumption against a historical baseline to highlight potentially unusual usage.
Anomaly Detection: Identify consumption patterns that may deserve further investigation.
What-If Savings: Explore potential savings under different energy-reduction assumptions.
AI-Powered Explanations: When the model integration is configured, generate understandable summaries and energy-saving recommendations.
Interactive Experience: Navigate analysis results through a connected, card-based interface.

WattWise treats unusual consumption as a potential opportunity for improvement, not automatic proof of energy waste.

How we built it

WattWise uses a modular architecture that separates numerical analysis from AI-generated explanations.

Frontend: React and Vite for the interactive user interface.
Backend: Python and FastAPI for API endpoints and application logic.
Data analysis: Pandas and NumPy for processing electricity consumption data.
Storage: superbase for saving analysis results.
AI integration: NVIDIA Nemotron through the NVIDIA API, when configured and available.

The Python calculation engine handles numerical analysis so that the core metrics remain reproducible. The AI model is used to explain the results rather than replace the underlying calculations.

Challenges we ran into

One challenge was designing an analysis process that can identify unusual energy consumption without incorrectly claiming that every anomaly is waste. We also needed to connect the frontend, backend, calculation engine, and AI service while keeping API credentials secure.

Separating deterministic calculations from AI-generated explanations helped us build a more transparent and testable system.

What's next

We want to expand WattWise with live smart-meter integrations, more sophisticated consumption forecasting, customizable electricity tariffs, and personalized energy-saving recommendations. Future versions could also compare energy patterns across days, weeks, and seasons.