Cartographic relevance sought in sustainable development may be found elsewhere
Cartographers at the International Cartographic Association's international conference have focused on sustainable development in an attempt to reassert their relevance – but it is in more novel forms of data processing and visualization they're likely to find greater professional relevance...
Outside of the mapping profession most people think cartography is dead, as they’ll tell you when you say you’re flying to Cape Town to attend a cartographic conference. The International Cartographic Association (ICA) annual conference’s focus on the profession’s relevance therefore seemed timely, and it took place under the theme “smart cartography for sustainable development”.
To be relevant, cartography needs to satisfy user needs by providing an interface with geospatial data that makes the underlying data usable, said Georg Gartner, the ICA’s newly elected president, in his keynote address.
Spatial knowledge graphs as a cartographic interface
Presentations about spatial knowledge graphs were among the most interesting, as they dealt with the usability of spatial data. Knowledge graphs model networks and relationships between data or entities. Spatial knowledge graphs include spatial entities in this network, and offer new ways to integrate, process, visualize, and manage spatial data.
By contextualizing the relations between entities, spatial knowledge graphs can be used to study the interaction between spatial layers and ontological information, explains José Pablo Ceballos Cantú from Vienna University of Technology. Furthermore, spatial knowledge graphs can fuse spatial and non-spatial data, allowing one to identify geographic relevance on a semantic basis. In this way, temporal, categorical and spatial nodes can simultaneously be represented as static nodes – meaning time and place can be analyzed in the same interface. This has implications for data management and analyses, such as identifying semantic relevance that might not be spatially obvious, including time-based relevance.
Antoni Moore, of the University of Otago, showed how he and the late Igor Drecki did exactly this by visualizing New Zealand’s national map collection in a spatial knowledge graph to identify inconsistencies and underrepresented geographies in the collection. They were also able to study themes of interest and identify nuances that were not apparent in the map archive.
By linking semantic data with map features, spatial data can be made usable in language models for language queries of spatial data – think Chat GPT for map data. As Wuhan University's Haijiang Xu explains, this kind of spatial reasoning would enable pattern-matching questions to be asked of spatial datasets, such as “where are features similar to building A?”.
