Information
- Publication Type: Master Thesis
- Workgroup(s)/Project(s): not specified
- Date: 2024
- Open Access: yes
- First Supervisor: Manuela Waldner
- Pages: 95
- Keywords: visual analytics, unstructured data, spatial organization, exploratory analysis, knowledge-assisted visual analytics, semantic interaction, knowledge externalization
Abstract
As not only the amount but also the complexity of data increases, there is a growing need to support humans in the analysis of data that is not structured in a way that can be easily interpreted by machines. So-called “knowledge-assisted visual analytics” (KAVA) tools aim to address these challenges by integrating the knowledge of the analyst into their system to support the analysis process.In this thesis, we investigate the spatial organization strategies that users employ when exploring unstructured data. We aim to characterize the types of strategies that users employ, how they change over time, and how we can use them to infer the users’ knowledge of the data. To answer these questions, we first conduct a user study in which the participants explore an image dataset on a multitouch tabletop interface imitating an analogue setting and externalize their findings into concept maps. We observe their organization strategies and analyse their methods in a mixed-methods approach, combining qualitative analysis of the participants’ interview statements with quantitative analysis of the interaction logs.We find that the participants’ spatial organization strategies can be characterized by four features: semantic clusters, type of layout, uncovering process, and reorganization of the data. While most participants prefer layouts that give them an overview of the data, only about half create semantic clusters (i.e., grouping similar images together). The participants also mostly uncovered all images — which were initially on a stack — in the task right away before externalizing their knowledge, and only a few reorganized the images. We further find that the participants generally did not change their organization strategies over time, and that the resulting spatial arrangements do not necessarily provide valuable insights into the users’ knowledge of the data.Finally, we discuss our findings and list the limitations of our study. As this thesis is embedded in a research project that aims to develop a tool for knowledge-assisted visual analytics, we discuss potential design implications for the development of such a tool.
Additional Files and Images
Weblinks
BibTeX
@mastersthesis{eitler-2024-sos,
title = "Spatial Organization Strategies in Exploratory Analysis of
Unstructured Data",
author = "Dominik Eitler",
year = "2024",
abstract = "As not only the amount but also the complexity of data
increases, there is a growing need to support humans in the
analysis of data that is not structured in a way that can be
easily interpreted by machines. So-called
“knowledge-assisted visual analytics” (KAVA) tools aim
to address these challenges by integrating the knowledge of
the analyst into their system to support the analysis
process.In this thesis, we investigate the spatial
organization strategies that users employ when exploring
unstructured data. We aim to characterize the types of
strategies that users employ, how they change over time, and
how we can use them to infer the users’ knowledge of the
data. To answer these questions, we first conduct a user
study in which the participants explore an image dataset on
a multitouch tabletop interface imitating an analogue
setting and externalize their findings into concept maps. We
observe their organization strategies and analyse their
methods in a mixed-methods approach, combining qualitative
analysis of the participants’ interview statements with
quantitative analysis of the interaction logs.We find that
the participants’ spatial organization strategies can be
characterized by four features: semantic clusters, type of
layout, uncovering process, and reorganization of the data.
While most participants prefer layouts that give them an
overview of the data, only about half create semantic
clusters (i.e., grouping similar images together). The
participants also mostly uncovered all images — which were
initially on a stack — in the task right away before
externalizing their knowledge, and only a few reorganized
the images. We further find that the participants generally
did not change their organization strategies over time, and
that the resulting spatial arrangements do not necessarily
provide valuable insights into the users’ knowledge of the
data.Finally, we discuss our findings and list the
limitations of our study. As this thesis is embedded in a
research project that aims to develop a tool for
knowledge-assisted visual analytics, we discuss potential
design implications for the development of such a tool.",
pages = "95",
address = "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
school = "Research Unit of Computer Graphics, Institute of Visual
Computing and Human-Centered Technology, Faculty of
Informatics, TU Wien",
keywords = "visual analytics, unstructured data, spatial organization,
exploratory analysis, knowledge-assisted visual analytics,
semantic interaction, knowledge externalization",
URL = "https://www.cg.tuwien.ac.at/research/publications/2024/eitler-2024-sos/",
}
thesis