Did you know the average US home emits nearly 9,000 pounds of CO2 a year just from electricity? Homeowners want to switch to clean energy, but they don't know where to start or if it makes financial sense for their specific home.
That's why we built Solwards.
Solwards is an intuitive web platform that takes the guesswork out of going solar. Users simply pull up their address and outline their roof on our interactive map. Our backend instantly crunches local solar irradiance data, the usable area considering the impact of obstructions like HVACs, buildings and trees, and the orientation of your roof to generate a personalized "Solar Score".
We also appeal to the pragmatists in every household owner and provide detailed fiscal data: approximate commitment, expected savings and ROI over 25 years.
Tech Stack
Backend Server: Python, Flask, DigitalOcean Gradient Platform Frontend: React.js, Vite
Google Maps API provides Precision Mapping. We use it to power our active interface. Providing users with a spatial visualization of the Earth for easy drawing ingesting the usable area from the roof.
NASA’s Power API gives us hyper-local solar irradiance data from latitude and longitude, allowing us to calculate the exact amount of sunlight hitting that specific geometry throughout the year.
US Energy Information Administration (EIA) API was used to pull localized energy rates that allow our backend to calculate approximate payback periods, immediate cost savings, and a 25-year ROI based on the user's location.
Digital Ocean's Gradient Platform: Using serverless inference functions, we generate a clean, human-readable summary of the metrics, giving users clear, actionable insights at a glance.
Solwards makes the invisible cost of energy visible, empowering homeowners to take their first step toward a carbon-neutral future. Let's guide the world towards solar.
Challenges we ran into
This was a massively parallel, agentic coding task where all of us learned to co-ordinate with each other when everyone has multiple agents working on many parts of the projects at once.
- The "Solar Score" Logic: Beyond Just Irradiance We quickly realized that "high sun" does not automatically mean "high value." A desert home might have a 10/10 irradiance profile but low local utility rates, making the ROI slower. Conversely, a home in a high-cost urban area with moderate sun might be a financial goldmine.
We engineered a weighted scoring algorithm that moves beyond simple weather data. Our engine calculates the "true potential" by factoring in:
Physical Nuances: Accounting for roof orientation (azimuth) and pitch.
Spatial Constraints: Subtracting "dead zones" caused by HVAC systems, chimneys, and vents which reduce the usable square footage.
Environmental Shading: Estimating the impact of nearby trees and buildings that cast "solar shadows."
- Solving the "Data Overload" Paradox Our backend is a powerhouse. It returns a massive JSON payload containing dozens of complex metrics, from 25-year ROI and net benefits to seasonal CO2 avoidance. During early testing, we found that flooding a user with this much raw data caused immediate cognitive fatigue. The user didn't know where to look first.
To balance technical depth with user accessibility, we implemented a Progressive Disclosure design pattern. Through multiple hurried iterations, we understood the importance of simplicity in design.
Level 1 (The Hook): A simplified dashboard focused on the "Solar Score" and the summary. This gives the user an immediate "Yes/No" signal.
Level 2 (The Deep Dive): For the power user, we built an interactive modal showing net cost, total energy, and CO2 impact. Another layer reveals the magnitude of daily variation and a monthly generation graph, allowing users to explore the seasonal science and the month-by-month variability of their energy harvest. We think this approach keeps the interface clean for beginners while remaining "feature-complete" for serious investors.
Log in or sign up for Devpost to join the conversation.