A comprehensive framework on multidimensional modeling

Abstract

In this paper we discuss what current multidimensional design approaches provide and which are their major flaws. Our contribution lays in a comprehensive framework that does not focus on how these approaches work but what they do provide for usage in real data warehouse projects. So that, we do not aim at comparing current approaches but set up a framework (based on four criteria: the role played by end-user requirements and data sources, the degree of automation achieved and the quality of the output produced) highlighting their drawbacks, and the need for further research on this area. © 2011 Springer-Verlag.