Leveraging Generative AI for Spatial Data Preparation

Preparing spatial data is a critical yet complex step in Geographic Information Systems (GIS), involving acquisition, cleaning, transformation, and integration. These processes often require significant manual effort and technical expertise. With the emergence of large language models (LLMs), spatial data workflows can now be partially automated. This paper explores the potential of LLMs to support data preparation in GIS through a literature-based analysis and a practical case study on mapping quantum computing centers. The study demonstrates how LLMs improve data quality, generate scripts, and assist in integrating heterogeneous sources. Key benefits include reduced processing time and increased consistency. Limitations, such as the need for human oversight and data availability, are also discussed. The findings position LLMs as valuable assistants in spatial data workflows.