DUET: Exploring Event Visualizations on Timelines

Abstract

Through innovations in sensors and computer vision, event-based temporal data collection has increased significantly in recent years. However, little, if any, research has focused on improving GUIs for these systems. Many applications use a linear timeline, leading to large downtime gaps when there are sparse events. In the meantime, linear timelines benefit from creating a natural sense of temporal perception. Our project seeks to find a balance between the natural temporal perception that linear timelines offer and improving the usability of visualizing event-based data using timelines. Through the exploration of design alternatives, we aim to evaluate different types of timeline visualization through which users can leverage the temporal perception from linear timelines and quickly locate events of interest. We conducted two studies to evaluate our interfaces, first a formative assessment study with specialists and then an experimental study. Our findings inform the design of future event-based timeline visualizations.