Exploring OLAP aggregates with hierarchical visualization techniques
Svetlana Mansmann, Marc H. Scholl · 2007
This paper presents an approach to exploring multidimen-sional data cubes with hierarchical visualization techniques. Analysts interact with data in a predominantly “drill-down” fashion, i.e. from coarse grained aggregates towards the de-sired level of detail. We suggest that visual hierarchies are adequate for mapping the multiscale nature of decomposi-tion as they preserve the results of the entire interaction. We introduce a class of visual structures called Enhanced Decomposition Tree. Every tree level is created by a disag-gregation step along a chosen dimension, the nodes contain the corresponding sub-aggregates arranged into a chart and the edges are labeled with their dimensional values. Various layouts are proposed to account for different analysis tasks. Data cubes are queried using a schema-based browser which presents dimensions by the hierarchies of their gran-ularity levels, thus offering an efficient way of generating hierarchical visualizations. Multiple data cubes may be ex-plored in parallel along their shared dimensions. The power of our approach is exemplified using a real-world study from the domain of academic administration.