Open Data Quality Dimensions and Metrics: State of the Art and Applied Use Cases

Abstract

While the economic benefit of open data is undeniable, its use as an asset in industrial processes is still a challenge. The lack of quality is indeed a typical argument for not leveraging open data. In fact, per the Data Office of the French government (ETALAB) in charge among others of the French open data initiative, only 9 out of the 34822 open datasets are tagged as reference datasets, that is supplied by certified publishers and which content can be reliable to be shared broadly, privately and publicly. Yet, no actual quality indicators are provided along with the metadata catalog of these 9 files.

What would then be the appropriate indicators for open data quality assessment, how would they differ from those used to assess DQ of traditional enterprise data? How can they be measured knowing the multiple reusability scenarios and how can they help users choose the datasets that best fit the purpose?

In this work-in-progress paper, we will answer these open data quality indicators questions and illustrate it with some case studies from the industry.