Implementing the Semantic Units Framework using Named Graphs and R2RML in a Knowledge Graph for the Biodiversity Exploratories

Tarek Al Mustafa, Lars Vogt, Tina Heger, Birgitta Koenig-Ries

Keywords: Knowledge Graphs, Semantic Units, Named Graphs, FAIR Principles, R2RML Mapping Language, Biodiversity Exploratories

TL;DR: We describe our implementation approach for the Semantic Units Framework using R2RML and Named Graphs in an OWL/RDF-based KG that contains publication and dataset metadata of the Biodiversity Exploratories.

Abstract:

Knowledge Graphs (KGs) can support many applications in the domains of ecology and biodiversity research, especially by facilitating synthesis efforts and the creation of FAIR datasets.
Semantic units have been proposed to organize a KG into semantically meaningful subgraphs using named graphs, to simplify cognitive interoperability for users, tackle common KG modeling and knowledge representation challenges, and facilitate reasoning over diverse logical frameworks, resulting in FAIR and CLEAR KGs.
However, no implementation beyond a small proof of concept and thus no evaluation of semantic units currently exists, which motivates our contributions:
We construct a KG for publication and dataset metadata of the Biodiversity Exploratories, a platform for functional biodiversity research in three study regions across Germany, and contribute the first implementation of semantic units on that graph.
We present relevant theoretical background on semantic units, detail our KG modeling methodology, and semantic unit implementation method using the R2RML mapping language and named graphs.
Finally, we list the findings from a competency question evaluation that compares SPARQL queries with and without semantic units and present the potential of semantic unit visualization to simplify the visual representation of KG content.
We conclude with lessons learned and avenues for future work based on our implementation.

Submission Number: 10