OnTeGra: Ontology-based access to Temporal and Graph Data

Formulating and evaluating queries can be challenging and resource-intensive, especially when dealing with data from heterogeneous, potentially incomplete, and unstructured sources. The Virtual Knowledge Graph (VKG) paradigm offers a solution by enabling users and applications to view disparate data sources as a unified knowledge graph, where nodes and edges are labelled with a familiar vocabulary. This approach has already proven effective in many application domains. However, several other areas have yet to benefit from it, as existing technologies do not meet their specific requirements. The OnTeGra project aims to address two major limitations: the lack of support for graph databases, which have gained significant popularity over the past decade, and for temporal data, which is ubiquitous. Our objective is to expand the formalisms and algorithms underlying the VKG paradigm to support graph databases in the form of property graphs and extensions of RDF, as well as to incorporate temporal data in simple data exchange formats with varying temporal granularity. With these extensions, users will be able to express and efficiently evaluate queries that involve complex navigation and pattern matching, temporal aggregations such as moving averages, trends, maxima and minima over temporal intervals, and traditional temporal operators. We will develop algorithms for answering queries over both materialized and virtual temporal graph data and establish a foundation for query optimization. Despite a decade of research, no existing system can yet query temporal VKGs. While there is extensive theoretical work in this area, the underlying challenges are complex, and practical implementations are still lacking. The key to achieving the necessary breakthrough will come from real-world applications. Two concrete use cases will guide the design of formalisms and the identification of useful practical fragments, which we will implement and evaluate in a prototype. This project is a collaborative effort between TU Wien, the Free University of Bozen-Bolzano, and the Virtual Vehicle Research GmbH. Together, the researchers from these three institutions contribute in-depth knowledge, expertise in VKG system development, and practical experience with real-world data and applications. This combined expertise will enable us to extend the benefits of VKGs to temporal and graph-based data.