Problem-oriented visualization of multi-dimensional data sets

The amount of data produced in the fields of engineering and science by measurements and simulations is steadily increasing. Scientific Visualisation has proven to be often the only reasonable approach to analyse this data to gain knowledge about the underlying processes and relations. However, the decision for the right visualisation methods is not easy at all. In general, the user must be an expert to generate an effective visualisation which considers the goal in mind and renders the characteristics of the given data set without producing artefacts and hereby misleading visualisations. This problem is of great impact since such wrong visualisations may lead to wrong conclusions. To support the user in the process of the generation of a visualisation it is necessary to provide guidance and tools which recommend visualisation methods and visualisation mappings, or even select an appropriate method. In this paper we discuss the various aspects of such visualisation support systems and present Vis-Wizz, a prototype of a problem-driven visualisation assistant.