Implementation of Metadata Repository based on Object-Oriented Model for Data Quality Assessment

Abstract

The paper considers the approach to organizing a metadata repository as one of the elements of the data quality assessment system. An object-oriented data model enriched with data quality checks is proposed for repository formation. A description of the key repository elements is given, along with an approach to linking data quality checks to objects in the data model. The article provides a method for storing the discussed metadata repository and special algorithms for its processing, including the "unpacking" algorithm for class attributes and the algorithm for determining the relevant data checks considering possible overrides. Based on the described theoretical propositions, a prototype of the metadata repository was implemented. The prototype was used to organize checks for assessing the data quality of personal data operator’s registry. In comparison with the implementation of a metadata repository based on a physical data model, the application of the approach described in this research results in a reduction of attribute and data quality check description by 23% and 27%, respectively, while maintaining the same quantity of executed checks. The investigated approach can be useful in practical tasks related to data quality analysis as a potential way to reduce the workload of data quality check management.