DQAF: Towards DQV-Based Dataset Quality Annotation Using the Web Annotation Data Model

Abstract

The W3C Data on the Web Best Practices Working Group is standardizing the Data Quality Vocabulary (DQV) for expressing data quality of Web-published datasets. As proposed in the DQV specification, quality annotations on datasets, one kind of quality information described using DQV, are achieved through Web annotations. Meanwhile, the W3C Web Annotation Working Group is creating a standard Web Annotation Data Model on the basis of the Open Annotation (OA) Data Model. Despite the significant progress in standardization, there is a lack of systematic research on Web tools for DQV-based dataset quality annotation. This paper therefore proposes a Dataset Quality Annotation Framework (DQAF) that provides annotating users with an interactive visual user interface, through which DQV-based dataset quality annotation can be readily achieved using the OA data model to produce machine-readable quality annotation data. We have implemented a proof-of-concept prototype of DQAF and conducted case study experiments with the prototype. The results indicate that DQAF is feasible and implementable, and annotating users can obtain a better understanding of and more intuitive interaction with the quality annotation data by means of the user interface.