Realizing embodied interaction for visual analytics through large displays

Visual analysts are engaged with the arduous task of scrutinizing increasingly larger datasets. Where conventional desktop displays are reaching their limits in terms of performance efficiency with large datasets, analysts can turn to larger displays. In a world of extensive multi-scale datasets, large high-resolution displays have the potential to show both more overview and detail for a given dataset than their smaller counterparts. In addition, people are able to use their embodied resources, such as spatial memory, proprioception, and optical flow to help them maintain orientation and improve performance times on analytic tasks when using larger displays. This paper looks at how physical navigation, physically interacting with large scale visualizations (e.g. walking, crouching, moving the head), affects user performance times on analytic tasks, such as finding patterns in geospatial data. The paper extends the space-scale diagram to take into account physical navigation and explains the theoretical repercussions. The paper then explains an empirical study performed for the purpose of further understanding how physical and virtual navigation affect performance times of tasks on varying size displays. In general, we found that large displays can decrease performance time of basic visualization tasks by more than ten times. In addition, we found overwhelming evidence from the empirical study that participants preferred physical navigation over virtual navigation (e.g. mouse interaction). Specifically, we found that for a number of tasks 100% of the participants chose to physically navigate physically moving to different areas on the display instead of using virtual navigation to manipulate the view on the display.