Empirical validation of metrics for conceptual models of data warehouses

Abstract

Data warehouses (DW), based on the multidimensional modeling, provide companies with huge historical information for the decision making process. As these DW's are crucial for companies in making decisions, their quality is absolutely critical. One of the main issues that influences their quality lays on the models (conceptual, logical and physical) we use to design them. In the last years, there have been several approaches to design DW's from the conceptual, logical and physical perspectives. However, from our point of view, there is a lack of more objective indicators (metrics) to guide the designer in accomplishing an outstanding model that allows us to guarantee the quality of these DW's. In this paper, we present a set of metrics to measure the quality of conceptual models for DW's. We have validated them through an empirical experiment performed by expert designers in DW's. Our experiment showed us that several of the proposed metrics seems to be practical indicators of the quality of conceptual models for DW's. © Springer-Verlag 2004.