Pointillist and Glyph-based Visualization of Nanoparticles in Formation

Abstract

In this paper we offer new, texture-based methods for the visualization of multivariate data. These methods aim to more effectively convey the results of calculations simulating the formation of nanoparticles in turbulent fl ows. In these simulations, an entire distribution of nanoparticles is computed at every point across a two-dimensional slice of the data space, for every time step. Previous visualization methods have relied on multiple separate images to convey summary statistics about the datasets, including mean diameter and standard deviation of particle sizes. We introduce new methods based on texture which aim to enable the integrated understanding of the entire distribution of values at each point across the domain in terms of both summary statistics at each point and particle counts for various sizes of particles. Pointillism is used to represent the data at each point across the data range as a high-resolution texture. Circular glyphs can also be used to form a more discrete, spot-based texture, in which different characteristics of the distribution are encoded in various features of the spots.