Methods for the Quality of Data in Databases

摘要

This paper contains possibilities to improve the quality of the (meta) data of the Central Relations Database (CRD) of a Dutch insurance company. This research provides methods and tools to measure and evaluate the quality of data (measure methods). Main causes of problems with the Quality of Data (QoD) are: the unclear, inconsistent and missing delivery of data and the absence of formal work procedures. This causes problems like false data in the database, data loaded at the wrong place, too many administrative activities, unclear conceptual models, difficulties in interpreting the meaning of data elements, etc. This means that procedures and methods have to be developed and applied to reach the desired QoD for both the user-organization and the data process-organization (corrective methods). Additional solutions are: introduction of formal documentation, development of uniform systems, replacement of the Data Dictionary System, enhancement of systems developers' responsibilities and creating a new function to store knowledge of employees. A general approach will be proposed.