EuroVA2020

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/2632904

Browse

Now showing 1 - 15 of 15

Dual Radial Set
(The Eurographics Association, 2020) Matkovic, Kresimir; Gracanin, Denis; Bardun, Matea; Splechtna, Rainer; Hauser, Helwig; Turkay, Cagatay and Vrotsou, Katerina
Set-typed data visualizations require novel interactive representations, especially when visualizing multiple set-typed data attributes. The challenge is how to effectively analyze relations between data elements from different set-typed attributes. We build on Radial Set view to support simultaneous visualization of two set-typed attributes. The main contributions include: Dual Radial Set view that supports simultaneous visualization of two groups of sets; an extension of Radial Set view that can display empty sets; and two new view configurations, the equal sector size and the relative-size scaled sectors. The two new view configurations also can be applied to the original Radial Set view. We conducted an informal evaluation using a movies data set as a case study. The evaluation results demonstrate the advantages of the proposed approach.
Enhanced Attribute-Based Explanations of Multidimensional Projections
(The Eurographics Association, 2020) Driel, Daan van; Zhai, Xiaorui; Tian, Zonglin; Telea, Alexandru; Turkay, Cagatay and Vrotsou, Katerina
Multidimensional projections (MPs) are established tools for exploring the structure of high-dimensional datasets to reveal groups of similar observations. For optimal usage, MPs can be augmented with mechanisms that explain what such points have in common that makes them similar. We extend the set of such explanatory instruments by two new techniques. First, we compute and encode the local dimensionality of the data in the projection, thereby showing areas where the MP can be well explained by a few latent variables. Secondly, we compute and display local attribute correlations, thereby helping the user to discover alternative explanations for the underlying phenomenon. We implement our explanatory tools using an image-based approach, which is efficient to compute, scales well visually for large and dense MP scatterplots, and can handle any projection technique. We demonstrate our approach using several datasets.
DualNetView: Dual Views for Visualizing the Dynamics of Networks
(The Eurographics Association, 2020) Pham, Vung; Nguyen, V. T. Ngan; Dang, Tommy; Turkay, Cagatay and Vrotsou, Katerina
The force-directed layout is a popular visual method for revealing network structures, such as clusters and important vertices. However, it is not capable of representing temporal patterns, such as how clusters/communities evolve. Dynamic network visualizations trade the overall structures for temporal relationships. In this paper, we present a dual view framework for capturing both overall structures and temporal patterns within networks. The linked supplemental views utilize the strengths of both visualization techniques to provide useful insights into the given networks. To demonstrate the usefulness of our proposed dual views, we provide three use cases of dynamic networks: computer networks communications, activities of suspicious processes in computer systems, and social networks.
EuroVa 2020: Frontmatter
(The Eurographics Association, 2020) Turkay, Cagatay; Vrotsou, Katerina; Turkay, Cagatay and Vrotsou, Katerina
An Exploratory Visual Analytics Tool for Multivariate Dynamic Networks
(The Eurographics Association, 2020) Boz, Hasan Alp; Bahrami, Mohsen; Suhara, Yoshihiko; Bozkaya, Burcin; Balcisoy, Selim; Turkay, Cagatay and Vrotsou, Katerina
Visualizing multivariate dynamic networks is a challenging task. The evolution of the dynamic network within the temporal axis must be depicted in conjunction with the associated multivariate attributes. In this paper, an exploratory visual analytics tool is proposed to display multivariate dynamic networks with spatial attributes. The proposed tool displays the distribution of multivariate temporal domain and network attributes in scattered views. Moreover, in order to expose the evolution of a single or a group of nodes in the dynamic network along the temporal axis, an egocentric approach is applied in which a node is represented with its neighborhood as an ego-network. This approach allows users to observe a node's surrounding environment along the temporal axis. On top of the traditional ego-network visualization methods, such as timelines, the proposed tool encodes ego-networks as feature vectors consisting of the domain and network attributes and projects them onto 2D views. As a result, the distance between projected ego-networks represents the dissimilarity across the temporal axis in a single view. The proposed tool is demonstrated with a real-world use case scenario on merchant networks obtained from a one-year-long credit card transactions.
Quality Metrics to Guide Visual Analysis of High Dimensional Genomics Data
(The Eurographics Association, 2020) Fernstad, Sara Johansson; Macquisten, Alexander; Berrington, Janet; Embleton, Nicholas; Stewart, Christopher; Turkay, Cagatay and Vrotsou, Katerina
Studies of genome sequenced data are increasingly common in many domains. Technological advances enable detection of hundreds of thousands of biological entities in samples, resulting in extremely high dimensional data. To enable exploration and understanding of such data, efficient visual analysis approaches are needed that take domain and data specific requirements into account. Based on a survey with bioscience experts, this paper suggests a categorisation and a set of quality metrics to identify patterns of interest, which can be used as guidance in visual analysis, as demonstrated in the paper.
Progressive Parameter Space Visualization for Task-Driven SAX Configuration
(The Eurographics Association, 2020) Loeschcke, Sebastian; Hogräfer, Marius; Schulz, Hans-Jörg; Turkay, Cagatay and Vrotsou, Katerina
As time series datasets are growing in size, data reduction approaches like PAA and SAX are used to keep them storable and analyzable. Yet, finding the right trade-off between data reduction and remaining utility of the data is a challenging problem. So far, it is either done in a user-driven way and offloaded to the analyst, or it is determined in a purely data-driven, automated way. None of these approaches take the analytic task to be performed on the reduced data into account. Hence, we propose a task-driven parametrization of PAA and SAX through a parameter space visualization that shows the difference of progressively running a given analytic computation on the original and on the reduced data for a representative set of data samples. We illustrate our approach in the context of climate analysis on weather data and exoplanet detection on light curve data.
Interactive Visualization of AI-based Speech Recognition Texts
(The Eurographics Association, 2020) Wu, Tsung Heng; Zhao, Ye; Amiruzzaman, Md; Turkay, Cagatay and Vrotsou, Katerina
Speech recognition technology has achieved impressive success recently with AI techniques of deep learning networks. Speechto- text tools are becoming prevalent in many social applications such as field surveys. However, the speech transcription results are far from perfection for direct use in these applications by domain scientists and practitioners, which prevents the users from fully leveraging the AI tools. In this paper, we show interactive visualization can play important roles in post-AI understanding, editing, and analysis of speech recognition results by presenting specified task characterization and case examples.
Congnostics: Visual Features for Doubly Time Series Plots
(The Eurographics Association, 2020) Nguyen, Bao Dien Quoc; Hewett, Rattikorn; Dang, Tommy; Turkay, Cagatay and Vrotsou, Katerina
In this paper, we propose an analytical approach to automatically extract visual features from doubly time series capturing the unusual associations which are not otherwise possible by investigating individual time series alone. We have extended the visual measures for 2D scatterplots, incorporated univariate time series analysis, and proposed new visual features for doubly time series plots. These measures are discussed and demonstrated via visual examples to clarify their implications and their effectiveness. The results show that distributions, trend, shape, noise, among other characteristics, can be used to uncover the latent features and events in temporal datasets.
A Window-based Approach for Mining Long Duration Event-sequences
(The Eurographics Association, 2020) Vrotsou, Katerina; Nordman, Aida; Turkay, Cagatay and Vrotsou, Katerina
This paper presents an interactive sequence mining approach for exploring long duration event-sequences and identifying interesting patterns within them. The approach extends previous work on exploratory sequence mining by using a sliding window to split the sequence prior to mining. Patterns are interactively grown and visualized through a tree representation, while a set of accompanying views allows for identified patterns to be explored in the context in which they occur. The approach is motivated and exemplified in the domain of air traffic control and, in particular, air traffic controller training.
SpatialRugs: Enhancing Spatial Awareness of Movement in Dense Pixel Visualizations
(The Eurographics Association, 2020) Buchmüller, Juri F.; Schlegel, Udo; Cakmak, Eren; Keim, Daniel A.; Dimara, Evanthia; Turkay, Cagatay and Vrotsou, Katerina
Compact visual summaries of spatio-temporal movement data often strive to express accurate positions of movers. We present SpatialRugs, a technique to enhance the spatial awareness of movements in dense pixel visualizations. SpatialRugs apply 2D colormaps to visualize location mapped to a juxtaposed display. We explore the effect of various colormaps discussing perceptual limitations and introduce a custom color-smoothing method to mitigate distorted patterns of collective movement behavior.
A Generic Model for Projection Alignment Applied to Neural Network Visualization
(The Eurographics Association, 2020) Cantareira, Gabriel Dias; Paulovich, Fernando V.; Turkay, Cagatay and Vrotsou, Katerina
Dimensionality reduction techniques are popular tools for the visualization of neural network models due to their ability to display hidden layer activations and aiding the understanding of how abstract representations are being formed. However, many techniques render poor results when used to compare multiple projections resulted from different feature sets, such as the outputs of different hidden layers or the outputs from different models processing the same data. This problem occurs due to the lack of an alignment factor to ensure that visual differences represent actual differences between the feature sets and not artifacts generated by the technique. In this paper, we propose a generic model to align multiple projections when visualizing different feature sets that can be applied to any gradient descent-based dimensionality reduction technique. We employ this model to generate a variant of the UMAP method and show the results of its application.
Learning and Teaching in Co-Adaptive Guidance for Mixed-Initiative Visual Analytics
(The Eurographics Association, 2020) Sperrle, Fabian; Jeitler, Astrik; Bernard, Jürgen; Keim, Daniel A.; El-Assady, Mennatallah; Turkay, Cagatay and Vrotsou, Katerina
Guidance processes in visual analytics applications often lack adaptivity. In this position paper, we contribute the concept of co-adaptive guidance, building on the principles of initiation and adaptation. We argue that both the user and the system adapt their data-, task- and user/system-models over time. Based on these principles, we propose reasoning about the guidance design space through introducing the concepts of learning and teaching that complement the existing dimension of implicit and explicit guidance, thus, deriving the four guidance dynamics user-teaching, system-teaching, user-learning, and system-learning. Finally, we classify current guidance approaches according to the dynamics, demonstrating their applicability to co-adaptive guidance.
Visual Analysis for Hospital Infection Control using a RNN Model
(The Eurographics Association, 2020) Müller, Martin; Petzold, Markus; Wunderlich, Marcel; Baumgartl, Tom; Höhn, Markus; Eichel, Vanessa; Mutters, Nico T.; Scheithauer, Simone; Marschollek, Michael; Landesberger, Tatiana von; Turkay, Cagatay and Vrotsou, Katerina
Bacteria and viruses are transmitted among patients in the hospital. Infection control experts develop strategies for infection control. Currently, this is done mostly manually, which is time-consuming and error-prone. Visual analysis approaches mainly focus disease spread on population level.We learn a RNN model for detection of potential infections, transmissions and infection factors. We present a novel interactive visual interface to explore the model results. Together with infection control experts, we apply our approach to real hospital data. The experts could identify factors for infections and derive infection control measures.
SepEx: Visual Analysis of Class Separation Measures
(The Eurographics Association, 2020) Bernard, Jürgen; Hutter, Marco; Zeppelzauer, Matthias; Sedlmair, Michael; Munzner, Tamara; Turkay, Cagatay and Vrotsou, Katerina
Class separation is an important concept in machine learning and visual analytics. However, the comparison of class separation for datasets with varying dimensionality is non-trivial, given a) the various possible structural characteristics of datasets and b) the plethora of separation measures that exist. Building upon recent findings in visualization research about the qualitative and quantitative evaluation of class separation for 2D dimensionally reduced data using scatterplots, this research addresses the visual analysis of class separation measures for high-dimensional data. We present SepEx, an interactive visualization approach for the assessment and comparison of class separation measures for multiple datasets. SepEx supports analysts with the comparison of multiple separation measures over many high-dimensional datasets, the effect of dimensionality reduction on measure outputs by supporting nD to 2D comparison, and the comparison of the effect of different dimensionality reduction methods on measure outputs. We demonstrate SepEx in a scenario on 100 two-class 5D datasets with a linearly increasing amount of separation between the classes, illustrating both similarities and nonlinearities across 11 measures.

Browse

Browsing EuroVA2020 by Issue Date

Results Per Page

Sort Options