Multidimensional Schemas Quality Assessment

A data warehouse is a database focused on decision making. It is built separately from the transactional (OLTP) databases of the enterprise, although it is partly fed from transactional data. Data warehouses are typically accessed by decision makers using OLAP tools, based on a specific, multidimensional representation of data. Considering the strategic importance of data warehouses, the quality of these systems is crucial. Moreover, since OLAP tool users directly access multidimensional schemas, multidimensional schemas quality is a key aspect of data warehouse quality. This paper focuses on the quality of multidimensional schemas. We present the underlying multidimensional model and the multidimensional schemas quality evaluation framework, which considers three views corresponding to three types of users: the specification view (data warehouse designer), the usage view (decision maker) and the implementation view (data warehouse developer). Concentrating on the specification view, we present and exemplify criteria and metrics pertaining to this view.