UC San Diego

Advances in technology now make it possible to capture detailed multimodal data about real-world everyday activity. Researchers have taken advantage of these advances to address questions about activity in more systematic and precise ways. Along with exciting opportunities to record data in ways that were not possible before, there are also analysis challenges due to the quantity and richness of the data. New tools and techniques are needed to address the challenges of analyzing dynamic multimodal activity data. I address the question of how such tools should be designed by developing an understanding of the data navigation patterns exhibited by researchers while they perform analysis. This thesis presents ChronoViz, a novel tool to support navigation, visualization, annotation, and analysis of multiple streams of time-coded data. ChronoViz supports analysis of data through the use of synchronized interactive visual representations of multiple data streams. Since visualizations are linked by time, each data stream can be used for navigation of the data set as a whole, and visualizations can flexibly be added, configured, and arranged to suit individual analysis needs. ChronoViz served three functions in this research. It enabled collaboration with researchers who collect observational time-based data, recorded the researchers' activity during use of ChronoViz for their data analysis, and provided a platform to enable design of interactive visualizations. Researchers used ChronoViz in their existing research projects, so the activity recorded was that of real analysis efforts. I describe patterns of navigation, how specific designs support those patterns, and identify ways to better support them in the future