Title: Uncertainty-Aware Visual Analytic Applications: From theory to usage


Name: Robin G. C. Maack


Phone: +49 631 205 3268




Project description:

The ever-increasing amount of data present in scientific research and industry already led to various data processing concepts and applications. Here, the results from sensors or simulations are being processed to extract essential information hidden in the data. Unfortunately, almost every real-world application suffers from uncertainty in the underlying data that is often ignored during data processing. Even when no uncertainty would be present in a dataset, uncertainty can also be introduced by the processing steps themselves, the mapping of data to a visualization, and the users’ perception of the representation provided by the application.




The Visual Analytics cycle of Keim et al. provides a starting point for my research, allowing users to process data by creating visualization and hypothesis that can be converted into insight. As a result, the insight influences the data creating a feedback loop that enhances the knowledge base in every iteration. Various sources of uncertainty can influence the components of this cycle that is not only specific to the component but also the application. Therefore, a new uncertainty-aware Visual Analytics cycle will be created that maps the uncertainty sources to the components, allowing to guide readers in their creation of Visual Analytics Applications that should include uncertainty. Previously published applications will be analyzed, evaluated, and improvements necessary to fully cover the new uncertainty-aware cycle design will be discussed.



Expected Results

Visual Analytics application creators that want to incorporate uncertainty can utilize a standard model that forms their application layout. This allows them to find sources of uncertainty to incorporate, visualization techniques that fit their data, and potential starting points for their solution. Additionally, already published applications show use cases of the uncertainty-aware Visual Analytics cycle to give users examples for their own developments.