Scatterplot Visualization of Hierarchically Clustered Data Points

Daniel Gruncl

Supervisor(s): Ing. Ladislav Čmolík, Ph.D.

Czech Technical University

Abstract: The focus of this paper is real-time visualization of and interaction with scatterplots with hundreds of thousands of data points, where the points are organized into a hierarchy of clusters. We present a technique that automatically selects a color palette for the clusters of the selected level. The level of the cluster hierarchy and the color palette are dynamically adjusted by zooming the scatterplot. Furthermore, the technique improves visibility of the displayed clusters by reducing occlusion of the overlapping clusters. We demonstrate the visualization technique using two real medical datasets containing 2D coordinates of hundreds of thousands points.
Keywords: Scientific Visualization
Full text:
Year: 2023