Introduction

SolScope is a sustainability-focused tool that evaluates parcels of land for their solar energy potential and provides AI-driven insights for solar energy deployment. By analyzing topography, solar irradiance, and proximity to power grids, and predicting future solar trends, our tool helps landowners, developers, and policymakers make data-driven, sustainable decisions.

To view and run the project for yourself, please head to this URL: https://github.com/Nedas-Jaronis/AIDaysHackathon2025?tab=readme-ov-file

This project was made by a collaborative team of developers who enjoy turning ambitious ideas into working products. Their contributions are listed below.

• Matthew Starks (Data Acquisition & Preprocessing): Handled API integration, data ingestion, cleaning, solar scoring algorithms, CSV compiling, and documentation/brand creation.

• Nedas Jaronis (Frontend + Machine Learning): Designed and built the user interface, including visualizations, built interactive AI heat maps and models, and designed the pipeline between front and backend.

• Kevin Duong (Backend): Managed backend processing, database integration, and data pipeline setup. 
  

Problem

How many prime solar sites are sitting unused simply because landowners and policymakers can’t see their potential?

Even though solar adoption is exploding at record rates, it is slowed down by one core issue: we aren’t clearly identifying the best locations for solar deployment. Landowners, developers, and state agencies lack simple, accessible tools that show:

• Where solar would produce the most energy

• Which parcels are worth investing in

• How factors such as geography, shading, or grid proximity affect viability

This has the following effects:

• Environmental: Valuable solar-ready land goes unused, slowing clean energy adoption and reducing potential carbon reduction.

• Economic: Landowners and developers miss out on profitable leasing and investing opportunities tied to solar-based projects.

• Policy/Community: Without clear insights, states and municipalities struggle to plan renewable infrastructure effectively, which slows down progress towards climate goals.

Data

According to the U.S. Energy Information Administration (EIA), solar adoption in the United States continues to grow overall. However, recent quarters show a noticeable slowdown, signaling a potential challenge for sustaining long-term renewable energy momentum [1].

Our web app focuses on the U.S. market and aims to accelerate clean energy adoption by identifying the most promising locations for solar development. Although the U.S. has the land, capital, and talent necessary to lead the way in clean energy, it still generates a smaller percentage of its electricity from solar power than other regions do.

We plan to accelerate adoption and support smarter renewable infrastructure decisions by providing actionable insights on where solar can thrive.

Project Overview

SolScope evaluates properties and regions using a suite of advanced analytics:

1. Solar Suitability Scoring: Calculates a single score based on terrain slope, price, area, annual solar irradiance, and proximity to power grid infrastructure.

2. AI Property Suggestions: Users may click on a property and locate similar properties with comparable attributes.

3. Renewable Energy Forecasting: Predicts whether a state or region will see an increase or decrease in renewable energy adoption in a future period of time (user selection) utilizing an AI model based on historical data.

4. Heat Map Visualization: Clusters of current solar deployment are displayed on an interactive map generated using AI, showing hotspots and areas underutilized for solar energy.

Our solution drives real impact by:

• Empowering landowners and developers find high-potential solar properties efficiently and make informed investment decisions.

• Supporting state and regional planning through clear, data-driven insights for sustainability and infrastructure development.

• Visualizing solar adoption trends to guide future renewable energy projects and accelerate clean-energy deployment.

Tech Stack

Overview

• Frontend: React.js (interactive maps, heat maps, AI suggestion interface)

• Backend: Python (data processing, scoring algorithms, AI prediction models)

• Database: SQLite for storing property and scoring data

APIs & Data Sources

• OpenStreetMap for grid distances

• National Solar Radiation Database (NSRDB) for solar irradiance

• Open Meteo Elevation for terrain elevation and slope

• US Energy Information Association for state renewable energy forecasting

Architecture Stack

If the image fails to load, please head to https://imgur.com/a/I4iA3iN.

Technical Deep-Dive

• AI Property Suggestions: The system identifies properties with comparable solar potential by standardizing multi-dimensional features—including geographic coordinates, solar irradiance (GHI/DNI), tilt, proximity to substations, solar score, and acreage—and computing Euclidean distances in this feature space. Using k-nearest neighbors, it ranks and returns the top-k properties most similar to a query, enabling data-driven recommendations for high-potential solar sites.

• Forecasting Renewable Trends: This pipeline predicts multi-year changes in renewable energy adoption using a combination of machine learning classifiers and regressors. It leverages historical energy data with engineered features including current percentages, year-over-year changes, and lagged values. A Random Forest classifier estimates the likelihood of renewable growth each year, while regressors predict precise percentages of renewable and non-renewable energy. The model iteratively forecasts year by year, updating features dynamically to produce a multi-year projection of renewable trends for each state, enabling data-driven energy planning and strategy.

• Renewable Energy Forecast Visualization: This pipeline extends multi-year renewable energy forecasting by combining historical state-level energy data with machine learning predictions (Random Forest classifier and regressors). It generates year-by-year projections for renewable and non-renewable percentages, dynamically updating features using lagged values and year-over-year changes. The module produces line plots overlaying historical data with forecasted trends, providing a clear visual representation of projected energy adoption for each state, enabling stakeholders to track and compare future energy scenarios.

• Calculating Sol Scores: Using data from 4 APIs and 6 weighted categories, we calculated ratings for each individual property by using historical solar irradiation and geospatial data.

• Integration: All scores, predictions, and visualizations update dynamically for real-time analysis.

• Full-Stack Application: Our system uses a three-tier architecture: the React frontend sends API requests to a FastAPI backend, the backend applies business logic and queries a SQLite database, and the results are returned as JSON for the UI to display.

Time Usage

This table summarizes how our 3-person team invested our time. Over the span of the hackathon, we dedicated roughly 100 total developer hours to completing this project.

Challenges

• Aligning multiple datasets (DEMS, irradiance, and renewable energy adoption) for consistency.

• Training AI models to give meaningful property suggestions.

• Finding slope values for a region based on property addresses.

• Fetching property addresses currently for sale using a non-paid method.

Lessons Learned:

• Preprocessing and normalization are crucial for AI accuracy.

• Communication is key when connecting different parts of the project together

• Defining roles prevents duplicate work from being done

• Geospatial data required more cleaning than expected

Accomplishments

• Implemented AI property suggestions for personalized solar insights.

• Built a renewable energy forecasting model for all states and regions in the United States.

• Developed interactive heat maps highlighting solar adoption clusters and underutilized areas.

• Successfully integrated multiple datasets to maximize sustainability impact.

• Learned how to integrate React, FastAPI, and SQLite with scikit-learn machine learning models.

Business/Monetization Strategy

Expenses: Ongoing expenses will primarily come from cloud compute, storage, and API usage, estimated at $50–$100 per month for a modest user base.

Revenue Model: We will use a freemium tier for landowners and individual users to buy and sell properties on the website. We will also offer subscription plans for developers, utilities, and policymakers for services such as batch predictions, advanced analytics, and solar irradiation overlays. In addition, we plan to partner with real estate platforms and earn commission-based revenue when users purchase land through referrals on our site.

Long-Term Vision: Partner with renewable energy agencies to guide national/state-level solar deployment initiatives, with solar sales companies to guide sales programs, and with real-estate companies to source land data.

Next Steps

Feature roadmap

• Expand data coverage to include more granular elevation, shading, and utility-grid layers.

• Add ROI projections per property for solar installation decisions.

• Create an “opportunities” tab to not only display properties for sale, but also properties that are off the market that are best suitable for solar development.

• Display ROI for installing solar panels on buildings and homes to access solar sales companies, homeowners, and business owners.

Service improvements

• Expand the dataset using a more reliable API to make searching for new properties in real-time possible. 

• Expand data layers (elevation, LiDAR, grid proximity) for more precise scoring.

Technical roadmap

• Containerize the application with Docker to ensure consistent deployments across environments.

• Deploy to Kubernetes for scalable, fault-tolerant infrastructure as usage grows.

• Deploy a stable cloud backend and open the platform for early pilot users.

• Set up CI/CD, cloud hosting, and monitoring for long-term stability

Business roadmap

• Launch a public beta for early landowners, solar installers, and sustainability groups to validate product-market fit.

• Contact real-estate platforms, solar agencies, and data providers for opportunities to expand reach and accelerate adoption.

• Develop premium features (batch evaluation, ROI calculators, advanced heat maps) to support subscription revenue.

• Expand our solar prediction platform into the Canadian market.

Frequently Asked Questions

Q: Can this tool predict solar potential for any U.S. property?
A: No, we used a sample dataset since we didn’t want to invest in premium property sale APIs that developers or investors utilize.

Q: How accurate are AI property suggestions?
A: They are based on similarity metrics from solar suitability features; accuracy improves as more data is added.

Q: How does this tool support sustainability?
A: By identifying optimal solar deployment sites, predicting adoption trends, and visualizing gaps, we maximize renewable energy generation and carbon reduction as well as promote the development of clean energy.

Sources

[1]: https://seia.org/research-resources/solar-market-insight-report-q3-2025/