On the Graph Structure of the Web of Data

This article describes how the Web of Data has emerged as the realization of a machine readable web relying on the resource description framework language as a way to provide richer semantics to datasets. While the web of data is based on similar principles as the original web, being interlinked in the principal mechanism to relate information, the differences in the structure of the information is evident. Several studies have analysed the graph structure of the web, yielding important insights that were used in relevant applications. However, those findings cannot be transposed to the Web of Data, due to fundamental differences in the production, link creation and usage. This article reports on a study of the graph structure of the Web of Data using methods and techniques from similar studies for the Web. Results show that the Web of Data also complies with the theory of the bow-tie. Other characteristics are the low distance between nodes or the closeness and degree centrality are low. Regarding the datasets, the biggest one is Open Data Euskadi but the one with more connections to other datasets is Dbpedia.