InDaQu: Enabling user-centered definition and exchange of consistency constraints for data cleaning

Severe data quality problems exist in most public health care systems and inconsistent data sets often occur. Consistency constraints can be used to define valid and invalid data. Existing solutions of such constraints like rule systems are often difficult to maintain, not human-readable, and of a bad quality like containing contradictory rules. With InDaQu we present an approach that allows domain experts to easily create and maintain consistency constraints using an introduced domain-specific language. These constraints are being stored in an ontology, which allows for an automated inconsistency detection in the defined rules themselves. We identified several scenarios in which consistency constraints can be interchanged and exchanged between different participants. The approach has been successfully evaluated in the cancer registry of Lower Saxony.