Thesis defense

Time and place

10th November 2015, at 13.15 p.m.
Auditorium 22.0.11, KUA, Njalsgade 120, building 22, 2300 Cph S.

Title

How does abundant display space support data analysis? Interaction techniques for information visualizations on large, high-resolution displays.

Abstract

This thesis explores information visualizations on large, high-resolution touch displays for analysis of massive amounts of data. The ever increasing rate at which data is collected about everything from peoples’ health, over organizations’ expenditures, to scientific experiments, necessitates new data analysis techniques. Information visualizations on large, high-resolution touch displays is a promising answer to these needs, and provide abundant display space for people to make sense of data. However, little is known about how to tailor interactive visualizations to abundant display space or about the benefits they might bring. The present thesis draws on the fields of human-computer interaction (HCI) and information visualization (InfoVis) to answer these questions.

This thesis is composed of four main studies. First, to explore how abundant display space influences data analysis, we conducted a workshop-based study with participants from a broad range of domains in which we use a whiteboard as a large display mock-up. We observed effects on for example visualizations’ lifespan, analysis provenance, and movement in relation to visualizations. In the second study, we focused on the benefits and drawbacks of combining movement and large display visualizations, by designing and evaluating such combinations. In the third study, we used insights from the workshop study and contextual inquiries with a group of healthcare data analysts, to design interaction techniques tailored to data exploration on large displays, and evaluate these in lab and deployment studies. The techniques allow people to create, configure, and freely lay out individual visualization views. In the fourth study, we specifically studied the manner in which visualizations shown with abundant display space might be related, to explore the variety of such relations and potential techniques to represent these. Finally, we synthesized the findings from these studies, and outlined future research directions.

Assessment committee

Heidrun Schumann, Professor at Institute for Computer Science, University of Rostock, Germany
Roy Ruddle, Reader at School of Computing, University of Leeds, UK
Sebastian Boring, Associate Professor at the University of Copenhagen, Denmark

Academic Supervisor

Kasper Hornbæk, Professor at Department of Computer Science, University of Copenhagen, Denmark

Electronic copy

Request an electronic copy directly from me