Solstice

Solstice uses real solar data and AI to make solar planning easy—letting anyone explore energy output, savings, and payback for any address through interactive maps and heat maps.

Comment

Track Chosen:

Renewables

Problem Statement:

Solar planning tools today are often overly technical and inaccessible to everyday users. They typically require engineering knowledge or data expertise, making it difficult for regular homeowners or city planners to estimate the benefits of solar energy in their specific area. We wanted to create a platform that makes this process intuitive, visual, and personalized—especially for communities like Atlanta, where solar potential can vary significantly from block to block.

Ideation & Development Process:

We started by identifying the gap between existing solar tools and what real users need—something simple, data-driven, and interactive. From there, we conceptualized Solstice, a web app that combines open solar data with AI to help users visualize and understand solar viability for their home, neighborhood, or city.

Our backend was built with Python using FastAPI, integrating two main external APIs: the NREL PVWatts API for solar energy production estimates and the Google Gemini API for conversational solar insights. The frontend was developed with React, using Leaflet for the interactive map interface and Recharts to create energy production and ROI visualizations.

During development, we faced challenges related to integrating and securing external APIs—specifically ensuring both address-based and coordinate-based solar queries worked reliably. Properly configuring PVWatts and managing API keys securely also required careful design choices.

Solution Proposed and Intended Impact:

Solstice lets users explore solar potential interactively. They can search for any address or simply click a location on a map to generate solar production, cost savings, and payback estimates. The app also creates large-scale heatmaps (like one of Atlanta) and smaller, detailed maps of specific areas, helping users compare solar potential across different points.

Additionally, users can chat with an AI assistant trained on real solar data to ask both personalized and general solar-related questions. This conversational layer makes the platform accessible to people without technical expertise.

Moving forward, we aim to incorporate detailed building features—like rooftop tilt and surface area—into calculations for more tailored recommendations. We also plan to enhance the chatbot to automatically generate full solar reports that users can export or share, empowering individuals and communities to make cleaner, data-backed energy decisions.

Share this project:

Updates