Technical Stack
React TypeScript TailwindCSS Leaflet Google Gemini AI NASA GIBS WMS Open-Meteo API
- NASA Satellite Data Integration
-
Integrated NASA's GIBS (Global Imagery Browse Services) WMS endpoint to fetch MODIS Terra NDVI (Normalised Difference Vegetation Index) 8-day composite imagery. The NDVI values quantify orchard vigour, providing a key metric for assessing plant health and bloom potential across the defined orchard polygon.
- Interactive Geospatial Mapping
-
Built an interactive mapping interface using Leaflet with custom drawing controls for orchard boundary definition. Users draw their exact orchard polygon on high-resolution satellite imagery, with automatic area calculation (in acres) using geodesic measurements and reverse geocoding for county identification.
- Hyper-Local Weather Integration
-
Implemented real-time 5-day weather forecasting using the Open-Meteo API, fetching temperature, wind speed, wind direction, and precipitation data for the orchard's centroid coordinates. A custom algorithm calculates optimal "Bee Flight Hours" based on temperature thresholds, wind conditions, and rain probability.
- AI-Powered Optimisation Engine
-
Developed a sophisticated prompt engineering system for Google Gemini AI that synthesises all quantitative inputs (NDVI, weather summary, orchard details) into a prescriptive "Pollination Plan". The AI acts as an expert agronomist, generating bloom phenology predictions, hive dosage recommendations, and detailed reasoning grounded in the input data.
- Actionable Operational Planning
-
The output includes a 5-day operational plan with daily recommendations based on predicted bee activity levels, a pollination schedule with phase-specific guidance (pre-bloom prep, hive placement, peak pollination, hive removal), and key metrics including recommended hive count and hives-per-acre ratio.
Challenge Response: BloomWatch
BloomBuzzer addresses NASA's "BloomWatch" challenge, which calls for tools that harness NASA Earth observation data to monitor and visualise plant blooming events and address specific vegetation monitoring, prediction, or management needs. The application demonstrates how satellite-derived NDVI can be translated into actionable insights for commercial agriculture.