UniVA: A Unified Interface for Visual Analytics
Final Report Abstract
The development of generic VA solutions that can be used in different application domains is challenging, as they require the integration of a wide range of functionality for visualization, interaction, and computational analysis. Instead of developing new "super applications", it is generally much more desirable and effective to combine and coordinate existing VA tools to provide the necessary functionality. In this project, we developed the novel concept of layered VA toolchains, which, in contrast to fixed VA frameworks, allow flexible coupling of independent VA tools. Our theoretical framework covers a coordination graph for modeling layered toolchains, complementary general data exchange mechanisms, and suitable unified UI ensembles. Together, they provide the basis for an effective, lightweight coordination of multiple tools and data in various VA scenarios. Based on our conceptual foundation, we implemented all our ideas in several software prototypes. These include a flexible editor to interactively create toolchains according to domain workflows and incorporate automated means for data exchange and transformations. Moreover, we designed a unified UI for effective display of all relevant information in a coordinated toolchain with several extensions to automated view arrangement and transitioning as well as annotation support. Together, our prototypes represent advances in the coupling of VA tools both on the data level and on the view level for a given analytical workflow. To provide access and encourage future use of our work, we made our solutions publicly available as open-source projects on GitHub. The repositories provided include a general description of our project, the AnyProc software for creating and executing toolchains, and the ReVize library for managing general data exchange. We demonstrated the generality and applicability of our solutions using two real-world scenarios together with domain experts: (i) the detection of cardiovascular abnormalities with our Health@Hand framework and (ii) the early detection of retinal abnormalities in diabetes mellitus with our unified UI. In both scenarios, we were able to show that our methods enable on-the-fly configuration of lightweight toolchains to work on real data. We also examined the critical care dataset MIMIC-III via a unified UI and applied our AnyProc editor to create a layered toolchain of multiple independent VA tools, which supported us in the data analysis. Executing the layered toolchains established a middle ground between a completely manual coordination of tools by domain experts on the one hand and the fully managed coordination of tools by tailored VA-systems on the other hand. The explorations made within this project are part of the broader research area of interactive visual data analysis. Our results specifically emphasize that suitable coordination means are crucial for working with multiple tools in interactive data analysis workflows in different application domains. Our work has also generated new ideas and enabled the establishment of ongoing research. In particular, we will continue our collaboration with ophthalmologists in a new joint research proposal to create a uniform interface for VA of interdisciplinary longitudinal eye study data. By bringing together the results of this project and novel techniques, we will be able to address not only the challenge of dealing with multiple tools and data, but also the added complexity of considering multiple therapeutic, diagnostic, and imaging modalities in the analysis workflows of different medical departments.
Publications
- “Health@Hand: A Visual Interface for eHealth Monitoring”. In: Proceedings of the IEEE Symposium on Computers and Communications (ISCC). 2019, pp. 1093–1096
L. Nonnemann, M. Haescher, M. Aehnelt, G. Bieber, H. Diener, and B. Urban
(See online at https://doi.org/10.1109/ISCC47284.2019.8969647) - “Lightweight Coordination of Multiple Independent Visual Analytics Tools”. In: Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (IVAPP). INSTICC. SciTePress, 2019, pp. 106–117
H. Schulz, M. Röhlig, L. Nonnemann, M. Aehnelt, H. Diener, B. Urban, and H. Schumann
(See online at https://doi.org/10.5220/0007571101060117) - Interactive Visual Data Analysis. AK Peters Visualization Series. CRC Press, 2020
C. Tominski and H. Schumann
(See online at https://doi.org/10.1201/9781315152707) - “A Characterization of Data Exchange between Visual Analytics Tools”. In: Proceedings of the International Conference Information Visualisation (IV). 2020, pp. 368–377
L. Nonnemann, H. Schumann, B. Urban, M. Aehnelt, and H.-J. Schulz
(See online at https://doi.org/10.1109/IV51561.2020.00066) - “A Layered Approach to Lightweight Toolchaining in Visual Analytics”. In: Computer Vision, Imaging and Computer Graphics Theory and Applications. Vol. 1182. Communications in Computer and Information Science. Springer, 2020, pp. 313–337
H. Schulz, M. Röhlig, L. Nonnemann, M. Hogräfer, M. Aehnelt, B. Urban, and H. Schumann
(See online at https://doi.org/10.1007/978-3-030-41590-7_13) - “Annotations in Different Steps of Visual Analytics”. In: Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (IVAPP). INSTICC. SciTePress, 2021, pp. 155–163
C. Schmidt, B. Grundel, H. Schumann, and P. Rosenthal
(See online at https://doi.org/10.5220/0010198001550163) - “Customizable Coordination of Independent Visual Analytics Tools”. In: Proceedings of the EuroVis Workshop on Visual Analytics (EuroVA). The Eurographics Association, 2021
L. Nonnemann, M. Hogräfer, H. Schumann, B. Urban, and H.-J. Schulz
(See online at https://doi.org/10.2312/eurova.20211094) - “Toward flexible visual analytics augmented through smooth display transitions”. In: Visual Informatics 5.3 (2021), pp. 28–38
C. Tominski, G. L. Andrienko, N. V. Andrienko, S. Bleisch, S. I. Fabrikant, E. Mayr, S. Miksch, M. Pohl, and A. Skupin
(See online at https://doi.org/10.1016/j.visinf.2021.06.004) - Towards a Unified User Interface for Visual Analysis of Retinal Data in Ophthalmology
M. Röhlig, L. Nonnemann, H.-J. Schulz, O. Stachs, and H. Schumann
(See online at https://doi.org/10.48550/arXiv.2301.01840) - “A Data-Driven Platform for the Coordination of Independent Visual Analytics Tools”. In: Computers and Graphics 106 (2022), pp. 152–160
L. Nonnemann, M. Hogräfer, M. Röhlig, H. Schumann, B. Urban, and H. Schulz
(See online at https://doi.org/10.1016/j.cag.2022.05.023)
