Solar Site Suitability Portal

Analyze, predict, and select the perfect spot for solar using real-time data and AI

Comment

⭐ Inspiration

Solar installation planning usually requires expert knowledge about irradiance, slope, elevation, and land conditions. I wanted to make this process simple and accessible. The idea was to build a tool where anyone can click on a map and instantly know whether a location (in India) is suitable for solar energy using real data and AI.

⚡ What it does

  • Fetches real solar irradiance from the NASA POWER API
  • Generates slope, elevation, and land-use scores synthetically
  • Runs a Random Forest ML model to classify the site
  • Computes a final suitability score using a weighted AHP-inspired formula
  • Visualizes results on an interactive map (Leaflet.js)
  • Displays analytics charts for irradiance, slope, elevation, and land score
  • Includes user authentication with signup and login

🛠️ How I built it

Node.js + MongoDB for authentication Flask (Python) for ML, NASA API integration, and suitability computation Leaflet.js for an interactive India map Chart.js for analytics dashboards TailwindCSS + Vanilla JS for the UI ML model trained on synthetic dataset based on heuristics AHP-style scoring function:

[ Score = 0.6 \cdot s_{\text{irr}} + 0.25 \cdot \left(1 - \frac{slope}{30}\right) + 0.15 \cdot landAI ] 🚧 Challenges I ran into

NASA API timeouts → solved using fallback synthetic irradiance No real labeled dataset → created synthetic training data Managing two servers (Node + Flask) without routing conflicts Dynamic chart generation required destroying and reinitializing charts Ensuring map interactions update all values smoothly

🏆 Accomplishments that I am proud of

Successfully integrated real NASA satellite data with ML predictions Built a fully functional map-based real-time analysis tool Implemented a hybrid model combining ML and AHP decision scoring Designed a clean analytics dashboard that updates based on user-selected cities Achieved a complete end-to-end workflow with two backends and multiple UIs

📚 What I learned

  • Integrating geospatial data with ML models
  • Designing synthetic datasets when real labels are unavailable
  • Handling cross-backend communication (Node ↔ Flask)
  • Building real-time map interactions with Leaflet
  • Applying AHP-style multi-criteria decision-making

🚀 What’s next for Solar Site Suitability Portal

Use real DEM (elevation/slope) data instead of synthetic Add rooftop segmentation using computer vision Provide ROI and cost-benefit analysis Deploy on cloud platforms for public use

Built With

Share this project:

Updates