A quality model for linked data exploration

Abstract

Linked (Open) Data (LD) offer the great opportunity to interconnect and share large amounts of data on a global scale, creating added value compared to data published via pure HTML. However, this enormous potential is not completely accessible. In fact, LD datasets are often affected by errors, inconsistencies, missing values and other quality issues that may lower their usage. Users are often not aware of the quality and characteristics of the LD datasets that they use for various and diverse tasks; thus they are not conscious of the effects that poor quality datasets may have on the results of their analyses. In this paper we present our initial results aimed to unleash LD usefulness, by providing a set of quality dimensions able to drive the selection and evaluation of LD sources. As a proof of concepts, we applied our model for assessing the quality of two LD datasets.