Towards a Mapping Framework for the Tenders Electronic Daily Standard Forms

Eugeniu Costetchi, Alexandros Vassiliades, Csongor Nyulas

Keywords: Knowledge Graph Generation, RDF Mapping Language (RML), Tenders Electronic Daily (TED), eProcurement Ontology (ePO), Public Procurement Data, Conceptual Mapping, Technical Mapping.

Abstract: Knowledge graphs are frequently built using declarative rules to bridge diverse data sources to a desired ontology and materialise them as RDF. The materialisation of the full knowledge graph may be a complex task when these data sources are extensive, making it unsuitable for an "on-demand" materialisation. In this paper, we present a methodology on how to map Public Procurement Data from the Tenders Electronic Daily website of the European Union by using RML, based on a innovative idea of mapping partitions. We map the aforementioned data into the eProcurement Ontology, which is a popular ontology when it comes to representing public procurement data. We also provide a method of evaluating the quality of the mapped data by using a mechanism that produces SPARQL queries based on the conceptual mapping of the Tenders Electronic Daily website data into the eProcurement Ontology. We then give an empirical evaluation over the quality of the produced data, and provide a detailed discussion on what the method presented in this paper has to offer.