GeoViz: A Multi-View Visualization Platform for Spatio-temporal Knowledge Graph
Jianping Zhou, Junhao Li, Guanjie Zheng, Yunqiang Zhu, Xinbing Wang, Chenghu Zhou
TL;DR
STKG visualization suffers from limited, single-view tools that hinder understanding of relationships, distributions, and evolution. The authors introduce GeoViz, an integrated, open-source platform offering three specialized views—knowledge tree, knowledge net, and knowledge map—with a no-code web interface for STKG exploration. The approach leverages hierarchical mapping, subgraph extraction, and LLM-based semantic similarity to reveal relationships and spatio-temporal patterns, plus a map and timeline for distribution observation. This work enhanced interpretability and usability for scientists and practitioners, with planned scalability to very large STKG and applications in domains like mountain hazards and smart-city planning.
Abstract
In this paper, we propose a multi-view visualization technology for spatio-temporal knowledge graph(STKG), which utilizes three distinct perspectives: knowledge tree, knowledge net, and knowledge map, to facilitate a comprehensive analysis of the STKG. The knowledge tree enables the visualization of hierarchical interrelation within the STKG, while the knowledge net elucidates semantic relationships among knowledge entities. Additionally, the knowledge map displays spatial and temporal distributions via spatial maps and time axes, respectively. Our visualization technology addresses the limitations inherent in single-view approaches and the deficiency of interaction in spatio-temporal perspectives evident in existing visualization methods. Moreover, we have encapsulated this technology within an integrated, open-source platform named GeoViz. A demo video of GeoViz can be accessed at https://github.com/JeremyChou28/GeoViz.