Graphing of E-Science Data with varying user requirements

Based on our experience in the Swiss Experiment, exploring experimental, scientific data is often done in a visual way. Starting from a global overview the users are zooming in on interesting events. In case of huge data volumes special data structures have to be introduced to provide fast and easy access to the data. Since it is hard to predict on how users will work with the data a generic approach requires self-adaptation of the required special data structures. In this paper we describe the underlying NP-hard problem and present several approaches to address the problem with varying properties. The approaches are illustrated with a small example and are evaluated with a synthetic data set and user queries.