Foundations of Visualization
An important part of work that is performed within the ISOVIS group is related to fundamental research that can be applied broadly. In consequence, this research is often part of concrete research projects and arises from specific visualization, computational, or empirical challenges in their context. In the following, we classify our fundamental research into three categories: human factors, fundamental algorithms, and taxonomies & design spaces.
Human Factors in Visualization and Visual Analytics
Humans are indispensable for the analytical process and the final decision making. One reason for this fact is that there are no or only insufficient computational models for many real-world problems, and humans are needed to perform or control many analysis tasks. Consequently, analysts and data scientists have to be supported to perceive, understand and share patterns or interrelations in large and complex data sets. As such, human factors are key to make this support as efficient and useful as possible and should play a central role in the design and evaluation of visualization tools.
Recent research examples: crowdsourcing for information visualization, the role of interaction in visual analytics, and supporting evaluations with EEG devices.
Fundamental and General Algorithms
Data mining in the widest sense is an integral part of human-centered visualization tools to make use of synergies between human analytical and computational capabilities. In addition, an efficient data management has to provide data mining and visualization components with appropriate data streams. Often, state-of-the-art algorithms do not work at all or simply not sufficiently when developing visualization systems. In those cases, the ISOVIS group develops own solutions that are mostly generalizable, i.e., can be applied in other visualization or data analytics contexts.
A recent research example is the development of a set of enhancements to the well-known Dynamic Time Warping algorithm, collectively called SlideDTW, for large streaming time series.
Taxonomies, Design Spaces, and Research Agendas
Many taxonomy and design space papers in the visualization areas were published in the past decades. They focus on a variety of aspects, such as tasks, interaction methods, visual designs/representations, or data types. Those works are very useful guide visualization experts and novices, but also for identifying gaps in the literature and open questions (i.e., providing a research agenda). A special case are large state-of-the-art reports or surveys that are nowadays often accompanied by online survey browsers.
Recent research examples are our surveys on sentiment visualization and text visualization, or the the identification of open problems in biological network visualization.