ENERGY · FULL-STACK WEB APP
SOLAR ORACLE TAPS REAL-TIME NREL DATA
Your location's solar potential, instantly — in dollars and CO₂, not kilowatt-hours per square meter
SEATTLE — Solar adoption is blocked less by cost than by uncertainty: most people have no idea how much sunlight their roof actually receives, or what it would mean in dollars saved or emissions avoided. Existing calculators are form-heavy and assume the user already knows their energy consumption in detail. SolarSave reduces the whole interaction to one input — a location.
Type a city, zip code, or full address. The server geocodes it through the OpenCage API, filtering out low-confidence results so solar data is never fetched for the wrong place, then queries the NREL Solar Resource API for real measured irradiance — not estimates or regional averages.
HOW IT WORKS
Raw irradiance means little to most people, so the app translates it: average peak sun hours per day, estimated annual electricity savings in dollars, and CO₂ offset in metric tons. A React Router flow separates search from results, keeping each screen focused.
The architecture runs two independent servers — a React 19 client and an Express 5 API — with both external API keys held server-side in environment variables, never exposed to the browser.
LESSONS FROM THE FLOOR
Chaining external APIs demanded error handling at every link: a low-confidence geocode had to be caught before it reached NREL, or the user would get a confident answer about the wrong location. The harder problem was translation — turning "4.8 kWh/m²/day" into units people actually care about proved to be the design challenge, not the data plumbing.
Source available → github.com/SNBest1/solarsave ↗
CONTINUED: Predictive speech cascade takes first place, Page A1 →