| dc.description.abstract | Recently, the Knowledge Graph (KG) based approach has become increasingly popular for integrating geospatial data. This popularity is due to the flexible nature of the graph data model that KGs have, and the ability to define precise data modelling through the means of ontologies. This allows for creating data models that represent such datasets. The same concept can be said for Virtual Knowledge Graphs. Virtual Knowledge Graphs (VKG) share the same foundational concepts as traditional KGs, but with a key difference: the VKG does not reside in a graph database. In traditional KG concepts, the data is stored in RDF graphs while a VKG makes it possible to let the data be stored in the data source itself only be present in query time. This ability is powerful for exploring large and unstructured datasets, such as geospatial data. However, using VKGs for geospatial data integration is still in its early stages. The ATKIS standard provides detailed, digital topographical data of the landscape and terrain of the Federal Republic of Germany. The most important component is the Digital Landscape Model (DLM), which provides data organised into various themes (e.g., Area, Settlement, Traffic). However, cross-analysis of these themes is challenging because they are not semantically linked. In this thesis, we identify these challenges associated with geospatial data integration of the ATKIS domain and propose a framework for integrating these datasets as a Virtual Knowledge Graph. | |
