Dense and Deep Geographic Virtual Knowledge Graphs for Visual Analysis

点击翻译按钮获取中文摘要

The D2G2 project has the goal to effectively integrate and analyze heterogeneous geodata sources by bridging two research fields – Virtual Knowledge Graphs (VKG) and Geovisual Analytics. D2G2 will develop a set of methodologies and software tools to create densely and deeply linked geospatial knowledge graphs, and it will offer a suite of intuitive visualizations and explainable analytical services over them. Recently, the Knowledge Graph (KG)-based approach has become increasingly popular for the integration of geodata, thanks to the flexibility of the graph data model and the possibility to define precise data semantics through ontologies. The GeoVKG paradigm enjoys the same benefits as classical KGs, and additionally allows us to “virtualize” existing geodata sources, e.g., relational databases. It avoids moving the data out of their original sources to materialize a possibly very large KG, and therefore is particularly useful for integrating large and complex datasets. Nevertheless, using VKGs for geodata integration is still in its infant stage for two main reasons. First, the VKG technologies have to be specialized and extended so as to capture geodata, by considering domain knowledge from the GIS community. Second, research covering the full lifecycle of GeoVKGs from construction to analysis is still largely missing. These challenges will be collaboratively addressed in D2G2 by two research groups from different disciplines, namely the Chair of Cartography at the TU Munich, led by Prof. Liqiu Meng, and the KRDB Research Centre at the Free University of Bozen-Bolzano, led by Prof. Diego Calvanese. The project will provide a complete solution covering the whole life-cycle from the construction and enrichment of GeoVKGs, to their visualization and analysis over them. Specifically, the D2G2 project will: (I) extend the VKG technology to construct KGs from common geodata formats and standards such as CityGML and ATKIS DLM; (II) enrich the constructed KGs to make them deeply and densely connected as a unified graph, by discovering co-located objects and similar themes in different subgraphs; (III) provide map-based visualizations of the VKGs for the efficient interpretation and navigation by the general public; (IV) develop a visual analysis framework over the VKGs to perform complex analytic tasks and explain the results. The developed methodology and software tools will be demonstrated in two real-world use cases. The first one deals with climate data in Bavaria, which is currently being explored in the Climate Change Cluster in Bavaria, while the second one addresses tourism and mobility data in South Tyrol managed by NOI Techpark in Bolzano. The two use cases have different geospatial and temporal locations and thematic data sources, but face comparable challenges in building, visualizing and exploring dense and deep VKGs and have the potential to provide transferrable insights for other cases.