Analyzing SPARQL Query Logs: A Study on Schema Coverage and Query Content

The advent of Knowledge Graphs (KGs) has revolutionized knowledge representation, enabling enhanced understanding, reasoning, and interpretation of complex data for both humans and machines. As a crucial tool for addressing real-world challenges, KGs rely heavily on SPARQL queries for data access and manipulation. Despite their extensive use, the alignment of these queries with the underlying KG schema is not well-understood. This paper introduces’SPARQL schema coverage’ as a novel measure to assess the extent to which SPARQL queries reflect the KGs’ content and structure. Utilizing Bio2RDF SPARQL logs as a case study, the paper reveals a SPARQL schema coverage of 98%, demonstrating a strong alignment between user queries and the KG schema, thereby highlighting high user engagement. This finding is significant for KG engineers in reshaping ontology and for triple store administrators in enhancing performance through targeted caching. The study addresses key research questions regarding the nature and extent of KG elements referred to in user queries and their coverage of the available data. This approach not only provides a new perspective in KG utilization but also aids in optimizing KG design and application, offering valuable insights for the future development and optimization of knowledge graphs.