Automating Geospatial Analysis Workflows Using ChatGPT-4
Published in Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems, 2024
Authors: Qianheng Zhang, Song Gao
Publication date: 2024/10/29 Book: Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems Pages: 715-716 PDF: Available from ACM
Description: The field of Geospatial Artificial Intelligence (GeoAI) has significantly impacted domain applications such as urban analytics, environmental monitoring, and disaster management. While powerful geoprocessing tools in geographic information systems (GIS) like ArcGIS Pro are available, automating these workflows with Python scripting using AI chatbots remains a challenge, especially for non-expert users. This study investigates whether ChatGPT-4 can automate GIS workflows by generating ArcPy functions based on structured instructions. We tested prompt engineering’s ability to help large language models (LLMs) understand spatial data and GIS workflows. The overall task success rate reaches 80.5%. It is a valid and easy to implement approach for domain scientists who want to use ArcPy to automate their workflows.
Keywords: GeoAI, ChatGPT-4, GIS, ArcPy, Workflow Automation