Hierarchy-driven Visual Exploration of Multidimensional Data Cubes

Abstract

Analysts interact with OLAP data in a predominantly “drill-down” fashion, i.e. gradually descending from a coarsely grained overview towards the desired level of detail. Analysis tools enable visual exploration as a sequence of navigation steps in the data cubes and their dimensional hierarchies. However, most state-of-the-art solutions are limited either in their capacity to handle complex multidimensional data or in the ability of their visual metaphors to provide an overview+details context. This work proposes an explorative framework for OLAP data based on a simple but powerful approach to analyzing data cubes of virtually arbitrary complexity. The data is queried using an intuitive navigation in which each dimension is represented by its hierarchy schema. Any granularity level can be dragged into the visualization to serve as an disaggregation axis. The results of the iterative exploration are mapped to a specified visualization technique. We favor hierarchical layouts for their natural ability to show step-wise decomposition of aggregate values. The power of the tool to support various application scenarios is demonstrated by presenting use cases from different domains and the visualization techniques suitable for solving specific analysis tasks.