Peek Brush: A High-Speed Lightweight Ad-Hoc Selection for Multiple Coordinated Views

Linking+Brushing is a proven concept to reveal relationships across multiple views. Defining complex selections, however, may involve a significant interaction overhead. This paper proposes Peek Brush, a point-brush that is designed to temporarily select and highlight items hovered by the user's mouse cursor. This enables quickly skimming through the data to identify relationships between different data projections within seconds. The Peek Brush serves the purpose of defining a starting point to a more focused inspection using brushes with higher complexity. In order to achieve rapid visual updates, we discuss acceleration techniques like preprocessing, threading, and layering. As a result, the Peek Brush is able to scale to datasets with millions of entries. A case study demonstrates how the Peek Brush minimizes the interaction effort required from the user. It delivers a quick overview and reduces the time needed for the initial visual analysis step from minutes to seconds.