Volume 57, Issue 1 - Special Issue: Visual Analytics / Daniel Keim, Tobias Schreck

In many applications, domain-specific entities are easily compared and categorized if they are represented as high-dimensional feature vectors. To detect object similarities and to quantify coherent groups, analysts often visualize the vectors directly, aiming to identify clusters visually. However, common visualizations for high-dimensional data often suffer from information loss, occlusions and visual clutter for large and noisy data. In this case, structure is misleading and false insights are derived. We use topological concepts to provide a structural view of the points. We analyze them in their original space and depict their clustering structure using intuitive landscapes. We describe the visual analysis process to define and simplify the structural view and to perform local analysis by linking individual features to other visualizations.

The number of visualizations being required for a complete view on data non-linearly grows with the number of data dimensions. Thus, relevant visualizations need to be filtered to guide the user during the visual search. A popular filter approach is the usage of quality metrics, which map a visual pattern to a real number. This way, visualizations that contain interesting patterns are automatically detected. Quality metrics are a useful tool in visual analysis, if they resemble the human perception. In this work we present a broad study to examine the relation between filtering relevant visualizations based on human perception versus quality metrics. For this, seven widely-used quality metrics were tested on five high-dimensional datasets, covering scatterplots, parallel coordinates, and radial visualizations. In total, 102 participants were available. The results of our studies show that quality metrics often work similar to the human perception. Interestingly, a subset of so-called Scagnostic measures does the best job.

A rising number of epidemiological studies apply imaging technologies. Images are not features themselves but provide raw data from which features are extracted. Different to other applications of analysis of medical images the data is analyzed statistically across the cohort. It results in unique requirements regarding the development of methods to efficiently integrate varying domain knowledge into the process without compromising comparability of results across subjects or time. Examples from two different strategies are presented and discussed.

In this article, we describe the concept of video visual analytics with a special focus on the reasoning process in the sensemaking loop. To illustrate this concept with real application scenarios, two visual analytics (VA) tools are discussed in detail that cover the sensemaking process: (i) for video surveillance, and (ii) for eye-tracking data analysis. Surveillance data (i) allow discussion of key VA topics such as browsing and playback, situational awareness, and the deduction of reasoning. Using example (ii) – eye tracking data from persons watching video – we review application features such as the spatio-temporal visualization along with clustering, and identification of attentional synchrony between participants. We examine how these features can support the VA process. Based on this, open challenges in video VA will be discussed.

Over the last decade, advances in high-throughput technologies have resulted in a flood of new biological data. Here, individual samples can extend up into terabyte size. While potential applications are broad, ranging from biotechnology to medical applications, the analysis of these datasets poses massive challenges. In order to make use of the produced terabytes of data, these datasets need to be integrated, need to be mapped onto existing biological knowledge, and need to be explored by experts. We present UniPAX and BiNA, a scalable system for the integration and analysis of high-throughput data (genomics, transcriptomics, proteomics, and metabolomics) in a network context. A central data warehouse holds the core dataset. A flexible middleware can execute custom queries on this dataset and communicate with our visual analytics tool BiNA, the Biological Network Analyzer. We demonstrate how the combination of these tools permits an efficient analysis of large-scale datasets for medical applications.

Social networking services have become a major channel for the digital society to share content, opinions, experiences on activities or events, as well as on products, services and brands. Evaluating digital feedback on the latter can be a valuable asset for companies seeking product and consumer insights. However, the analysis of short, noisy, fragmented, and often subjective textual data still remains a challenge. Typically, the human analyst needs to be actively involved during extraction and modeling to resolve ambiguities that will inevitable arise in such data and to put the model into context. This paper proposes a visual analytics approach that enables a first intuition and exploration of topics appearing in the text corpus, and facilitates the interactive-iterative refinement of the overall topic model describing the stream of tweets. A second contribution is the discussion of efficient graph community detection algorithms to extract initial topics as the starting point of interactive analysis that complement approaches such as LDA. The applicability and utility of the proposed approach is shown for a real-world use case: the analysis of product insights and topic-driven social networks analysis for a specific product line for an international hair styling and cosmetics company.

My research interest lies at the interaction between human-computer interaction (HCI) and geoinformatics. I am interested in developing new methods and novel user interfaces to navigate through spatial information. This article will give a brief overview on my past and current research topics and streams. Generally speaking, geography is playing an increasingly important role in computer science and also in the field of HCI ranging from social computing to natural user interfaces (NUIs). At the same time, research in geography has focused more and more on technology-mediated interaction with spatiotemporal phenomena. By bridging the two fields, my aim is to exploit this fruitful intersection between those two and develop, design and evaluate user interfaces that help people to solve their daily tasks more enjoyable and effectively.

A university of applied sciences focuses on both application and science. Before becoming a professor at a university of applied sciences, I got in touch with both worlds: the purely scientific one, doing research in the field of algorithm engineering, and the `real' one, working in large software development projects. After making these experiences, I am convinced that the approach of a university of applied sciences is a very good one.

Clinical research and routine care documentation have traditionally been performed in a dual-source approach, where data is collected redundantly and maintained in separated systems. The generic x4T single-source process removes redundant steps and is adoptable to specific types of studies. In support of the various processes, we have designed and implemented an architecture called x4T based on CDISC ODM. The software has been deployed in three different use cases containing up to 3976 subjects until now and allows real-time access to live data from routine care documentation systems.