Description
Data visualization is one of the major issues of database research. OLAP a decision support technology, is clearly in the center of this effort. Thus far, visualization has not been incorporated in the abstraction levels of DBMS architecture (conceptual, logical, physical); neither has it been formally treated in this context. In this paper we start by reconsidering the separation of the aforementioned abstraction levels to take visualization into consideration. Then, we present the Cube Presentation Model (CPM), a novel presentational model for OLAP screens. The proposal lies on the fundamental idea of separating the logical part of a data cube computation from the presentational part of the client tool. Then, CPM can be naturally mapped on the Table Lens, which is an advanced visualization technique from the Human-Computer Interaction area, particularly tailored for cross-tab reports. Based on the particularities of Table Lens, we propose automated proactive support to the user for the interaction with an OLAP screen. Finally, we discuss implementation and usage issues in the context of an academic prototype system (CubeView) that we have implemented.