Understanding Data Quality in a Data Warehouse

Abstract

Data quality has long been recognised as a key factor in the success of data warehousing. Attempts to define data quality typically provide lists of quality dimensions that are vaguely defined, overlapping and not based in sound underlying theory. One promising approach is to define quality dimensions in ontological foundations. This paper presents an alternative approach that uses semiotic theory to develop a framework for understanding data (the content) and metadata (the structure) quality in a data warehouse. Goals of data and metadata quality are separated from the means of achieving them and ways of measuring them. Existing data and metadata quality dimensions are analysed and placed within the framework. The framework provides researchers and practitioners with a sound, theoretically-based set of quality goals, means and measures which will support further work in the development of data and metadata quality guidelines, evaluation procedures and empirical studies of data and metadata quality in practice.