Collaborative Visualization : Designing and evaluating systems for co-located work

Abstract: This thesis investigates new ways of using information visualization to support collaboration in co-located work. To study this phenomenon, Multiple Viewer Display Environments (MVDEs) with independent views have been applied to present information such that all viewers at the same time and in the same display can see correct views of 3D models, see correctly oriented text and see different parts and aspects of information in each view. Several prototypes have been developed either as proof of new conceptual designs or to evaluate particular research questions. These prototypes have been used to investigate general properties that apply to co-located collaborative visualizations. A prototype system to keep track of the viewpoints and information in the independent views was implemented on MVDE hardware to support discussions on future command and control environments and to provide the necessary framework for conducting empirical studies (Paper II). Another prototype, the in situ tomographic display, was developed to support presentation of spatial 3D data (e.g., temperature or airflow) in 2D views in situ with working environments (Paper III). In addition to the visualization systems, a technique for high precision pen-based interaction in rear-projection display environments - the PixelActiveSurface – was developed (Papers IV and V). The empirical studies evaluate how new forms of visualization in MVDEs with independent views affect the way information is perceived and can be shared in collaboration. The conclusion is that multiple independent views can provide more effective and efficient visualization when the following conditions are met: text is oriented towards the viewer (Paper VI), different aspects of information are coordinated between different views of the same display (Paper VIII) and correct views of 3D models are used to compare ordinal information and relations in spatial data (Paper VII). However, for the techniques to support co-located work efficiently, it is necessary that the type of work and the task to be solved are first properly analyzed and understood (Papers VII and IX).