A visualization method for domain knowledge graphs considering spatial scale transformations
The node-link graph,a classic method for visualizing knowledge graphs using force-oriented layouts,is widely applied across various fields due to its simplicity and clarity.However,when visualizing large-scale knowledge graphs,traditional methods often produce cluttered results that fail to emphasize the core content and spatial distribution patterns.To address these limitations,this study proposes a new method for domain knowledge graph visualization considering spatial scale transformations.By designing a screening model with multiple constraint rules and integrating knowledge graphs with geographic maps,this method generates a more organized distribution of nodes and links,leading to significantly improved visualization outcomes.The domain knowledge graph visualization method considers spatial scale transformations and consists of two key processes.First,all entities and relationships within the knowledge graph are passed through a screening model,where they are successively filtered based on spatial,temporal,attribute weight,and topic classification constraints.This process continues until the threshold conditions,determined by the map scale,are met.Second,the filtered entities and relationships are used to reconstruct the knowledge graph for visualization.By utilizing the spatial location information of entities,the method establishes a mapping between knowledge graph nodes and geographic points,allowing for spatial superposition on the map.The feasibility of this method is verified by a case study of the typhoon public opinion knowledge graph,and the visualization effect is compared with traditional methods in terms of spatial representation and multi-scale representation.Compared to the traditional node-link graph method with a force-directed layout,the knowledge graph visualization generated by the method considering spatial scale transformations offers significant advantages.In terms of spatial representation,the layout produced by this method more accurately reflects the real-world spatial distribution patterns of knowledge networks.It effectively aggregates entities with spatial characteristics,which closely aligns with actual geographic space.In terms of multi-scale representation,the number of nodes and links is properly controlled across different spatial scales.The visualization transitions smoothly as the spatial scale changes,maintaining clarity and reducing visual clutter.The domain knowledge graph visualization method considering spatial scale transformations enhances the traditional node-link graph by reducing visual clutter from the large numbers of nodes and links while emphasizing the core content and spatial distribution of the knowledge graph across different scales.This significantly improves both user experience and knowledge acquisition.In the case of the typhoon public opinion knowledge graph,the method accurately represents the real-world distribution of knowledge networks and dynamically filters irrelevant information based on the observation perspective,demonstrating its effectiveness.This method offers valuable insights into the integration of knowledge graphs with geographic data and the spatial representation of geographic knowledge.
knowledge graphforce-directed layoutspatial scale transformationsmulti-scale representationgeographic information
NETL
NSTL
万方数据
