In this paper we address the issue of conceptual modeling of data used in multidimensional analysis. We view the problem from the end-user point of view and we describe a set of requirements for the conceptual modeling of realworld OLAP scenarios. Based on those requirements we then define a new conceptual model that intends to capture the static properties of the involved information. In its definition we use a minimal set of well-understood OLAP concepts like dimensions, levels, hierarchies, measures and cubes. The central concept of the model is the Multidimensional Aggregation Cube (MAC), which gives a broad and flexible definition to the notion of a multidimensional cube. We evaluate our model against other existing multidimensional models and show that MAC offers a unique combination of modeling skills. Our main contribution is the definition of the basic concepts of our model; although the set of requirements and the evaluation of all related models against those requirements r...
Aris Tsois, Nikos Karayannidis, Timos K. Sellis