A Spatiotemporal Knowledge Graph Construction and Management System

Abstract. To fulfill the current lack of scalable and efficient software systems for constructing and managing spatiotemporal knowledge graphs, this research develops an integrated system that supports full lifecycle knowledge management, multimodal visualization, and the construction of both generic and domain-specific knowledge graphs. By collaboratively building, storing, and updating large-scale spatiotemporal knowledge graphs with hundreds of millions of triples, the system systematically resolves four key difficulties: First, it improves construction efficiency through human-machine collaboration and streamlined workflows. Second, with the help of an integrated knowledge update and completion engine, it enables dynamic updates of instances, entities, relationships, and other knowledge graph elements. Third, the system incorporates review and quality assessment mechanisms to ensure the reliability and consistency. Lastly, it offers multi-layered, multi-view visualization for intuitively displaying spatial distribution and temporal evolution. The system not only enables the engineering implementation of large-scale spatiotemporal knowledge graphs but also provides a closed-loop management framework, delivering a comprehensive solution for constructing, updating, quality-controlling, and providing complex spatiotemporal knowledge applications.