Abstract
The present disclosure proposes a system to address the challenges faced by professionals in fields like sustainability, urban planning, and clean energy development when analyzing geospatial data. Traditional methods for deriving actionable insights from disparate geospatial data sources are often complex, time-consuming, and require specialized expertise. This document describes a system and technique that leverages Large Foundational Models, such as Large Language Model (LLM) agents to automate geospatial workflows across multiple datasets from varied sources. The solution employs an LLM-assisted code generation and execution loop within a persistent execution environment, enabling efficient and simplified geospatial data analysis.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Sieniek, Marcin; Ponda, Sameera; Jung, Gia; Gupta, Ananya; Murphy, Gearoid; Merrikh-Yazdi, Danusch; Johnston, Renee; and Blohmé, Patrik, "Stateful LLM Agent for Geospatial Data Analysis", Technical Disclosure Commons, (June 01, 2025)
https://www.tdcommons.org/dpubs_series/8177