3D metaphoric information visualization

In the Information Age the difficult part is not producing or even having access to information. Our society has become increasingly rich in information and new analysis tools that enable users to make sense of it are demanded. Visualization goes beyond using computer graphics to communicate information; visualization enables us to think. This thesis studies using three-dimensional (3D) metaphoric representations for visualizing large volumes of abstract data. Specifically, it focuses four main challenges in 3D information visualization: mapping the information, visualizing it, navigating the representations, and managing dynamics. A key research problem in the visualization of abstract information, since it has no natural and obvious physical representation, is to discover visual mappings for representing the information and to understand what analysis tasks they support. We present a framework and a strategy for automatically mapping abstract information onto 3D metaphoric virtual worlds. We give several examples of implemented visual mappings that illustrate and validate our strategy. The fact that the information mapping is done automatically allows us to create different visualizations on the fly. Furthermore, since our mapping strategy is based on metadata, our mapping engine is virtually application domain independent. We took advantage of both features to develop multiple views for different domain applications and research their usefulness in information visualization. Now that available graphics computing power is not a hindering factor anymore, navigation is frequently seen as the weakest link in 3D virtual worlds. We introduce a novel concept of navigation -- metaphor-aware navigation -- and build on that concept to design and implement an integrated helped-navigation system that extends the traditional navigation systems present in most 3D viewers. We provide examples to illustrate our helped-navigation mechanisms, namely the metaphor-aware navigation. The visualization of dynamic data is still a seldom explored issue in information visualization. In this thesis, we investigate the impact of real-time visualization of dynamic data and we propose strategies to cope with data dynamics across the entire pipeline of canonical information-visualization systems. The impact of data dynamics on the visualizations is attacked with more detail and we devise constraint propagation strategies to minimize the disruptive effects the dynamics may have on the stability of the virtual world.