Electrigrid

Logo

Inspiration

As our cities grow and the demand for electricity surges, the threat of grid instability and rolling blackouts has become a very real challenge. We realized there is a massive disconnect between individual household energy habits and macro-level grid health. We were inspired to build Electrigrid to bridge this gap: creating a platform that not only empowers individuals to understand and optimize their personal consumption but also provides civic administrators with the macro-level data needed to predict and prevent blackouts.

What it does

Electrigrid is a dual-purpose smart energy management web application:

For the Individual User: A personalized dashboard tracks utility and device-level consumption in real-time. Users can monitor usage spikes, remotely control connected devices (like turning off the HVAC or lights), and receive tailored, actionable energy-saving tips from an integrated AI Recommendations Panel.

For the City Administrator: A secure admin panel provides aggregated, anonymized data on neighborhood and community usage. It features a predictive Blackout Possibility Panel that maps high-risk zones and calculates grid strain, allowing administrators to make proactive decisions before an outage occurs.

How we built it

We prioritized a highly responsive, modern architecture to handle real-time data seamlessly.

Frontend: We used ReactJS built with Vite for rapid performance. For the design system, we implemented a custom, minimalist color palette blending an Off-White canvas with Prussian Blue accents to convey trust and clarity. We used Framer Motion for smooth, intelligent page transitions and micro-interactions (like tactile device toggles).

Backend: We built a robust API using Node.js and Express.js, utilizing WebSockets to handle the low-latency remote device control.

Data & AI: The user-facing AI recommendations are powered by an LLM that analyzes the user's JSON consumption data to generate conversational tips. The admin blackout predictions run on a machine learning model that evaluates real-time load versus historical grid capacity.

Challenges we ran into

Handling high-frequency, real-time data from simulated IoT devices without bottlenecking the React frontend was a significant hurdle. We had to carefully manage our component lifecycles and state to ensure the data visualizations updated smoothly. Additionally, tuning the prompt engineering for the AI Recommendations Panel took several iterations to ensure the outputs were concise, accurate, and actually useful to the user, rather than generic advice.

Accomplishments that we're proud of

We are proud of the remote device control, the immediate visual feedback and smooth animations make the dashboard feel like a premium, native smart-home application.

What we learned

We deepened our understanding of real-time bidirectional communication using WebSockets and Express.js. On the design front, we learned how to balance dense data visualization with white space, ensuring that both a casual user and an advanced city administrator can read their respective dashboards at a glance.

What's next for Electrigrid

The immediate next step is integrating live smart-home APIs (such as Google Nest or SmartThings) to transition from simulated devices to real-world hardware control. For the admin side, we want to implement an automated "Load Shedding" communication protocol that can instantly send SMS alerts to users in high-risk zones, asking them to voluntarily reduce consumption during peak hours.