Towards Configurable Composite Data Quality Assessment

The growing availability of data over the last decades has given rise to a number of successful technologies, ranging from data collection and storage infrastructures to hardware and software tools for efficient computation of analytics. This context, in principle, places a great demand on data quality. As a matter of fact, experience has shown that the open Web and other platforms hosting user-generated content or real-time data can provide little quality control at content production time. To address these challenges, our aim is to provide a general and configurable model for assessing data quality supporting task composition. In particular, we introduce a model characterized along the notion of matching, illustrating the issues that can be addressed by this approach with a concrete case study. We also identify and discuss challenges to be addressed in future research to strengthen this idea.