1. ¹University of Konstanz, Konstanz, Germany
2. ²Stanford University, Palo Alto, CA, U.S.A.
3. ³University of Konstanz, Konstanz, Germany

Correspondence: Jörn Schneidewind, Computer Science Institute, Universität Konstanz Box D78, Universitätsstr. 10, 78457 Konstanz, Germany. Tel: +49 7531 88 3077; fax: +49 7531 88 3062; E-mail: schneide@inf.uni-konstanz.de

Received 23 June 2006; Revised 31 July 2006; Accepted 9 September 2006.

Abstract

During the last two decades, a wide variety of advanced methods for the visual exploration of large data sets have been proposed. For most of these techniques user interaction has become a crucial element, since there are many situations in which users or analysts have to select the right parameter settings from among many in order to construct insightful visualizations. The right choice of input parameters is essential, since suboptimal parameter settings or the investigation of irrelevant data dimensions make the exploration process more time consuming and may result in wrong conclusions. But finding the right parameters is often a tedious process and it becomes almost impossible for an analyst to find an optimal parameter setting manually because of the volume and complexity of today's data sets. Therefore, we propose a novel approach for automatically determining meaningful parameter- and attribute settings based on the combined analysis of the data space and the resulting visualizations with respect to a given task. Our technique automatically analyzes pixel images resulting from visualizations created from diverse parameter mappings and ranks them according to the potential value for the user. This allows a more effective and more efficient visual data analysis process, since the attribute/parameter space is reduced to meaningful selections and thus the analyst obtains faster insight into the data. Real-world applications are provided to show the benefit of the proposed approach.