@phdthesis{Mayer2024,
title = {From Exploratory to Explanatory Interactive Visualization of Spatio-Temporal Conflict Data},
author = {Benedikt Mayer},
year = {2024},
date = {2024-09-24},
abstract = {Violent conflicts cause human suffering and devastation worldwide. Conflict research studies the causes, dynamics, and impact of such conflicts, as well as ways to prevent them. Interactive data visualization can support conflict researchers in exploring corresponding data. However, so far, barely any studies have been conducted analyzing how to provide this kind of exploratory support in the field of conflict research.
In contrast, interactive data visualization is already commonly used in online stories reporting on conflicts, as it can help to explain space- and time-dependent developments more clearly. However, there is little structured research of how such datadriven visual storytelling can be performed for stories with a spatio-temporal context, like those about conflicts.
Exploratory and explanatory approaches can also be combined, allowing users to get a deeper understanding of the conveyed information. While this approach is more commonly used to target broad audiences in the context of science communication, it can also be used to communicate scientific information to researchers. However, little research has been performed in this direction as well.
To address the listed issues, we collaborated with a conflict researcher to conduct three main studies. They are positioned along a spectrum between exploration and explanation to answer the question: “How can interactive visualization be used to support the exploration and explanation of spatio-temporal conflict data?”
In the first study, we investigated how to design an application that allows conflict researchers to explore the results of a process in which multiple conflict data sets are integrated into one holistic set. We abstracted the underlying domain problems and derived a workflow and application design to solve the identified problems. An evaluation involving five conflict researchers showed that the application helped them to better understand and validate the results of the data integration. However, it also showed that even experts require a considerable amount of explanation to properly use such exploratory applications.
In the second study, we investigated how to combine exploration and explanation to communicate the workings of a scientific method to conflict researchers, based on a visual data story. We
distilled the story creation process and evaluated the story with eight conflict researchers. It showed that they accepted the storytelling approach well, and their feedback allowed us to derive a set of guidelines for performing such exploratory explanation for experts.
In the third study, we collected 130 data stories with a spatio-temporal context to investigate how visual storytelling was applied in them for explaining the underlying messages. To do so, we merged and adapted three existing design spaces and used them to classify the stories based on which storytelling techniques they employed. We analyzed the resulting data, identifying various ways to combine the storytelling techniques as well as trends that occurred over the years, e.g., towards more easily consumable stories, and towards updating and extending stories over time.
Overall, our studies showed several promising directions for how interactive visualization can support the exploration and explanation of spatio-temporal conflict data. However, it also became clear that more research is necessary. Accordingly, it should be investigated how to make exploratory applications more accessible through explanation, how to make the process of creating spatio-temporal data stories more efficient, and what the role of interaction in explanatory applications can be.},
keywords = {},
pubstate = {published},
tppubtype = {phdthesis}
}