GeoPredict
Project Overview
University
Year
About
Standard AI models often analyse locations as isolated data points, overlooking how nearby places influence one another and sometimes giving misleadingly confident results. GeoPredict is an open-source toolkit that incorporates geographic relationships into machine-learning models. It guides users through spatial analysis, suitable model selection, location-aware testing, prediction and uncertainty mapping, revealing where results are reliable and where errors may occur. By making spatial AI more transparent and accessible, the platform could improve decisions in urban planning, housing, public health, transport, energy and environmental monitoring.
Team
Rui Deng
Rui Deng is a PhD candidate in Geospatial Data Science at the University of Glasgow, holding degrees from the Chinese Academy of Sciences and Shandong University. His research focuses on GeoAI, geospatial tabular data mining, and vector representation learning, with a emphasis on adapting Transformer architectures and foundation models for spatial data. He has published first-author papers in premier venues including AAAI and IJGIS.