geospatial-ai-assistant-platform
nazaryansurenn-amb/geospatial-ai-assistant-platform
Parcel-level geospatial decision support with a tool-using LLM assistant (OpenAI function calling over deterministic Python tools), a Sentinel-2 / ERA5-Land pipeline with 18 spectral indices, and a React + MapLibre vector-tile map. Case study: ~44,000 cadastral parcels in Armenia. Code only.
Summary
A comprehensive geospatial AI platform for parcel-level land analysis in Armenia, combining a tool-using LLM assistant (OpenAI function calling) with deterministic Python analytics, a Sentinel-2/ERA5-Land data pipeline, and a React/MapLibre frontend. The system provides decision support for land activity, use type, history, potential, and consolidation across ~44,000 cadastral parcels.