Defense: D. Abramov (CM) - Interactive Analysis Tools for Visualizing the Universe

This thesis introduces three novel interactive astronomy data visualization tools for exploring and analyzing information from telescope observations and cosmological simulations. The amount of data produced by modern telescopes and simulations is tremendous and can be challenging for experts and non-experts to explore comprehensively without the help of accessible, interactive techniques and tools. Given the size and complexity of these data, it often requires advanced technical expertise to access, query, view, and analyze the information. Furthermore, working with large-scale simulations requires powerful computing resources that may not be accessible on basic consumer hardware. The projects described in this thesis leverage advances in consumer graphics hardware and browser-based rendering to provide lightweight, accessible tools to the astronomy community that are designed to simplify analytic workflows, reduce cognitive load and the need for users to code, and contextualize the data meaningfully. In Chapter 1, the introduction, a brief history of astronomy visualization methods from the stone age to now, is preceded by a description of the three interactive browser-based visualization tools introduced in this thesis: IGM-Vis (chapter 2) contextualizes the impact of galaxies in observational absorption spectra through the introduction of "skewer sightlines"; CosmoVis (chapter 3) is a cosmological simulation volume and particle visualizer that allows users to place "skewer sightlines" in the volume and produce absorption spectra and column densities; and finally, a temporal animation extension to CosmoVis (chapter 4) that allows for exploring the evolution of cosmological simulations. For each project, the scientific motivations and background are detailed for a visualization audience and contextualized with related software tools. We identify astronomy-based analysis tasks that each tool solves, along with details of the design and implementation of the systems as they relate to enabling the tasks. Finally, a set of scientific use cases, as well as user and performance evaluations, are provided.

Event Host: David Abramov, Ph.D. Candidate, Computational Media

Advisor: Angus Forbes