Publication Date

2023

Document Type

Dissertation/Thesis

First Advisor

Papka, Michael E.

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Computer Science

Abstract

Many real-world networks are multivariate in nature and also evolve over time. Such a network whose nodes and/or edges along with their associated attributes, evolve over time is called A dynamic multivariate network (DMVN). For instance, in the scholarly community, researchers often collaborate with others. These collaborations have various attributes, such as scientific domain and number of publications, as well as they change over time. At any given point in time, the collaborations among researchers can be modeled as a network resulting in a collaboration network. When the changes in collaborations and associated attributes are also included in the network, it results in a dynamic collaboration network, a type of DMVN. Exploring such kinds of networks is of great significance to the information visualization community but also challenging due to the presence of dynamic and multivariate nature at the same time. Hence, visualization of DMVNs focuses on the challenge of representing the evolution of relationships between entities and attributes on the entities and/or edges in a readable and scalable manner. Most existing techniques focus on either the dynamic nature of the topology or the multivariate nature of the network, lacking joint analysis of both characteristics.

This dissertation contributes an in-depth analysis of DMVN visualization techniques. We approach it first by organizing the design space of DMVN visualization techniques to understand the state of the art and identify new research opportunities, followed by three key contributions: 1) A taxonomy to classify DMVN visualization techniques; 2) Novel technique to improve the stability of layout when visualizing the structural changes in DMVNs; and 3) An empirical study that evaluates the support of two common visualization techniques for performing various multivariate network tasks: node-link diagrams and list view-based representations.

Extent

174 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS