本研究通过覆盖多种应用场景，旨在评估当前最先进的大语言模型（LLM）在生成地理处理工作流及解决地理空间问题方案方面的能力，并揭示此类模型在空间数据科学自动化进程中日益增强的作用。为构建基准测试任务，我们综合整理了2025年之前公开可获取的GIS教程（例如Esri Learn平台、高校实验指导材料）中的地理空间问题求解实例。

By addressing a variety of scenarios, this research aims to evaluate how well the state-of-the-art LLMs can generate geoprocessing workflows and solutions for geospatial problems while highlighting the evolving role of these models in the automation of spatial data science. To construct the benchmark tasks, we curated geospatial problem-solving examples from a combination of publicly available GIS tutorials (e.g., Esri Learn platform, university lab instructions) prior to the year 2025.