Search

Visual Analytics integrates the outstanding capabilities of humans in terms of visual information exploration with the enormous processing power of computers to form a powerful knowledge discovery environment. In other words, Visual Analytics is the science of analytical reasoning facilitated by interactive interfaces and captures the information discovery process keeping the human in the loop. Process Mining aims to extract information and knowledge from event logs to discover, monitor, and improve processes in a variety of application domains. Such event data or traces of activities often possess data quality issues as well as exhibit unexpected behavior and complex relations. Consequently, before and during the implementation of interactive (automated or semi-automated) analysis methods, such as Process Mining algorithms, the analyst needs to explore, investigate, and understand the data at hand in order to decide which analysis methods might be appropriate. The combination of interactive visual data analyses and exploration with Process Mining algorithms makes complex information structures more comprehensible and facilitates new insights. In this talk, I will illustrate the concepts of Visual Analytics, how Visual Analytics combined with Process Mining techniques supports to extract more insights from complex event data, and elaborate about the challenges and opportunities for analyzing process data with Visual Analytics methods. Various examples will illustrate what has been achieved so far and show possible future directions and challenges.

A temporal graph stores and reflects temporal information associated with its entities and relationships. Such graphs can be utilized to model a broad variety of problems in a multitude of domains. Re-searchers from different fields of expertise are increasingly applying graph visualization and analysis to explore unknown phenomena, complex emerging structures, and changes occurring over time in their data. While several empirical studies evaluate the benefits and drawbacks of different network representations, visualizing the temporal dimension in graphs still presents an open challenge. In this paper we propose an exploratory user study with the aim of evaluating different combinations of graph representations, namely node-link and adjacency matrix, and temporal encodings, such as superimposition, juxtaposition and animation, on typical temporal tasks. The study participants expressed positive feedback toward matrix representations, with generally quicker and more accurate responses than with the node-link representation.

Caused by a newly discovered coronavirus, COVID-19 is an infectious disease easily transmitted between people through close contacts that had exponential global growth in 2020 and became, in a very short time, a major health, and economic global issue. Real-world data concerning the spread of the disease was quickly made available by different global institutions and resulted in many works involving data visualizations and prediction models. In this paper, (1) we discuss the problem, data aspects, and challenges of COVID-19 data analysis; (2) We propose a Visual Analytics approach (called COVis) combining different temporal aspects of COVID-19 data with the output of a predictive model. This combination supports the estimation of the spread of the disease in different scenarios and allows correlating and monitoring the virus development in relation to different government response events; (3) We evaluate the approach with two domain experts to support the understanding of how our system can facilitate journalistic investigation tasks and (4) we discuss future works and a possible generalization of our solution.

Investment planning requires knowledge of the financial landscape on a large scale, both in terms of geo-spatial and industry sector distribution. There is plenty of data available, but it is scattered across heterogeneous sources (newspapers, open data, etc.), which makes it difficult for financial analysts to understand the big picture. In this paper, we present Sabrina, a financial data analysis and visualization approach that incorporates a pipeline for the generation of firm-to-firm financial transaction networks. The pipeline is capable of fusing the ground truth on individual firms in a region with (incremental) domain knowledge on general macroscopic aspects of the economy. Sabrina unites these heterogeneous data sources within a uniform visual interface that enables the visual analysis process. In a user study with three domain experts, we illustrate the usefulness of Sabrina, which eases their analysis process.

Visual Analytics (VA) aims to combine the strengths of humans and computers for effective data analysis. In this endeavor, humans’ tacit knowledge from prior experience is an important asset that can be leveraged by both human and computer to improve the analytic process. While VA environments are starting to include features to formalize, store, and utilize such knowledge, the mechanisms and degree in which these environments integrate explicit knowledge varies widely. Additionally, this important class of VA environments has never been elaborated on by existing work on VA theory. This paper proposes a conceptual model of Knowledge-assisted VA conceptually grounded on the visualization model by van Wijk. We apply the model to describe various examples of knowledge-assisted VA from the literature and elaborate on three of them in finer detail. Moreover, we illustrate the utilization of the model to compare different design alternatives and to evaluate existing approaches with respect to their use of knowledge. Finally, the model can inspire designers to generate novel VA environments using explicit knowledge effectively.

Recognizing activities has become increasingly relevant in many application domains, such as security or ambient assisted living. To handle different scenarios, the underlying automated algorithms are configured using multiple input parameters. However, the influence and interplay of these parameters is often not clear, making exhaustive evaluations necessary. On this account, we propose a visual analytics approach to supporting users in understanding the complex relationships among parameters, recognized activities, and associated accuracies. First, representative parameter settings are determined. Then, the respective output is computed and statistically analyzed to assess parameters' influence in general. Finally, visualizing the parameter settings along with the activities provides overview and allows to investigate the computed results in detail. Coordinated interaction helps to explore dependencies, compare different settings, and examine individual activities. By integrating automated, visual, and interactive means users can select parameter values that meet desired quality criteria. We demonstrate the application of our solution in a use case with realistic complexity, involving a study of human protagonists in daily living with respect to hundreds of parameter settings.

Missing values are a problem in many real world applications, for example failing sensor measurements. For further analysis these missing values need to be imputed. Thus, imputation of such missing values is important in a wide range of applications. We propose a visually and statistically guided imputation approach, that allows applying different imputation techniques to estimate the missing values as well as evaluating and fine tuning the imputation by visual guidance. In our approach we include additional visual information about uncertainty and employ the cyclic structure of time inherent in the data. Including this cyclic structure enables visually judging the adequateness of the estimated values with respect to the uncertainty/error boundaries and according to the patterns of the neighbouring time points in linear and cyclic (e.g., the months of the year) time.

One of the primary concerns of financial institutions is to guarantee security and legitimacy in their services. Being able to detect and avoid fraudulent schemes also enhances the credibility of these institutions. Currently, fraud detection approaches still lack Visual Analytics techniques. We propose a Visual Analytics process that tackles the main challenges in the area of fraud detection.

Large multivariate time-oriented networks have been gaining an increasing relevance in different domains. In order to support Visual Analytics processes on this kind of data, appropriate storage and retrieval methods are needed that take into account the scale, dimensionality, and in particular the complex nature of time. We introduce TimeGraph, a data management framework consisting of a data model and two levels of abstraction. TimeGraph captures both the topology of networks and the inherent structure of time into a property graph data structure, and transparently handles them by graph-based operations. TimeGraph aims to be an expressive, easy-to-use and extensible framework, enabling data reduction by selection and aggregation over both the temporal and the topological properties of data, to foster interactive visualization and analysis.

Reconstructing processes from measurements of multiple sensors over time is an important task in many application domains. For the reconstruction, these multivariate time-series can be automatically processed. However, the outcomes of automated algorithms often vary in quality and show strong parameter dependencies, making manual inspections and adjustments of the results necessary. We propose a visual analysis approach to support the user in understanding parameters' influences on these results. With our approach the user can identify and select parameter settings that meet certain quality criteria. The proposed visual and interactive design helps to identify relationships and temporal patterns, supports subsequent decision making, and promotes higher accuracy as well as confidence in the results.

The research field of Process Mining deals with the extraction of information from event logs. Since large amounts of event data are generated every day, analyzing these often unstructured event logs poses a research challenge. To cope with the complexity of the data and the associated mining tasks, appropriate visualizations and interactive means are needed. We present our early prototype for the visual exploration of event sequences and patterns. Our approach combines (1) a visual representation of event sequences emphasizing recurring event patterns, (2) automated pattern mining methods, as well as (3) interactive means for exploration. We provide first steps to support browsing event logs in an interactive environment and to facilitate the inspection of recurring pattern locations within the context of the surrounding events.

The visualization and analysis of dynamic social networks are challenging problems, demanding the simultaneous consideration of relational and temporal aspects. In order to follow the evolution of a network over time, we need to detect not only which nodes and which links change and when these changes occur, but also the impact they have on their neighbourhood and on the overall relational structure. Aiming to enhance the perception of structural changes at both the micro and the macro level, we introduce the change centrality metric. This novel metric, as well as a set of further metrics we derive from it, enable the pair wise comparison of subsequent states of an evolving network in a discrete-time domain. Demonstrating their exploitation to enrich visualizations, we show how these change metrics support the visual analysis of network dynamics.

The Eclipse Java debugger uses an indented list to view arrays at runtime. This visualization provides limited insight into the array. Also, it is cumbersome and time-consuming to search for certain values at an unknown index. We present a new Eclipse plug in for visualizing large arrays and collections while debugging Java programs. The plug in provides three views to visualize the data. These views are designed to support different tasks more efficiently. A tabular view gives detailed information about the elements in the array, such as the value of their field variables. A line chart aims to depict the values of a numerical field over the array. Lastly, bar charts and histograms show how the values of a field are distributed. We show how these views can be used to explore linear data structures and hashes from the Collections Framework. The plug in features tight integration with the Eclipse IDE, and is freely available as an open-source project. Developers' feedback confirmed the utility of the plug in to explore large arrays in real-world scenarios.

In contrast to stepping, tracing is a debugging technique that does not suspend the execution. This technique is more suitable for debugging programs whose correctness is compromised by the suspension of execution. In this work we present a tool for visually tracing Java programs in Eclipse. Trace point hits are collected on a per-instance basis. This enables finding out which trace points were hit for which objects at which time. The interactive visualization provides detailed information about the hits such as thread, stack trace, and assigned values. We implemented the tool as an Eclipse plug in that integrates with other features of Eclipse Java debugger. In an informal evaluation, developers appreciated the utility of our method as a solution in the middle between full tracing and stop-and-go debugging. They suggested scenarios in which our tool can help them in debugging and understanding their programs.

To make informed decisions, an expert has to reason with multi-dimensional, heterogeneous data and analysis results of these. Items in such datasets are typically represented by features. However, as argued in cognitive science, features do not yield an optimal space for human reasoning. In fact, humans tend to organize complex information in terms of prototypes or known cases rather than in absolute terms. When confronted with unknown data items, humans assess them in terms of similarity to these prototypical elements. Interestingly, an analogues similarity-to-prototype approach, where prototypes are taken from the data, has been successfully applied in machine learning. Combining such a machine learning approach with human prototypical reasoning in a Visual Analytics context requires to integrate similarity-based classification with interactive visualizations. To that end, the data prototypes should be visually represented to trigger direct associations to cases familiar to the domain experts. In this paper, we propose a set of highly interactive visualizations to explore data and classification results in terms of dissimilarities to visually represented prototypes. We argue that this approach not only supports human reasoning processes, but is also suitable to enhance understanding of heterogeneous data. The proposed framework is applied to a risk assessment case study in Forensic Psychiatry.

It is assumed that approximately one third of severe car accidents are related to drowsiness. Warning systems such as the Mercedes Benz Attention Assist try to tackle this problem by analyzing the driving style. Previous work investigated the estimation of measures (features) from lane data that correlate well with impaired driving. Unfortunately, these features require a lane-tracking camera, which is not available in many cars. Furthermore, the lane data signals are often affected from missing road markings, bad sight etc. Some lane-based features such as LANEDEV or ZIGZAGS do not require the absolute distance to the lane markings, but only depend on the lateral deviation within the lane. Our idea is to exploit odometric data (yaw rate and vehicle speed) to estimate this measure. The vehicle trajectory is a composition of the lurching between lane markings and the disturbing road curvature. Thus, we remove this curvature by a filter since its frequency is lower than the vehicle deviation. We compare the correlation between features based on lane data and odometric data as well as their relationship with sleepiness. An excerpt of the Attention Assist database with 294 drives and over 76 000 km is used. We show that some lane-based features can be approximated well. The zero-crossing rate (LATPOSZCR) performs even better than its lane-based pendant.

A model predictive control scheme for handling prioritized objectives in a plant under fault influence is presented. The approach is based on the lexicographic multiobjective optimization and uses a model predictive controller with a l 1 norm as performance criterion. The effectivity of the scheme is shown on a reverse osmosis plant under different fault cases like actuator saturation and stuck valves. The obtained results are satisfactory which motivates the continuity in the studies of this kind of fault-tolerant control.

In this paper we present a method to compensate the image motion of a monocular camera on a moving vehicle in order to detect obstacles. Due to the camera motion, the road surface induces a characteristic image motion between two camera shots. The motion of the camera is determined by the use of odometric data received from the CAN-bus, and the position and orientation of the road is continuously estimated with camera self-calibration. This all leads to a motion field which is predicted based on homography. To prevent the drawbacks of the real camera perspective, different virtual camera perspectives are presented in combination with motion compensation. Possible virtual perspectives are the bird's eye view and image rectification. In addition, a non-linear camera model is used which does not limit the range of obstacle detection to a certain distance and efficiently uses the available image information.

In this paper we present a method to calibrate the extrinsic parameters of a monocular camera on a moving vehicle. The method is based on a homography between two camera shots. Therefore, only the road surface has to be visible in the pair of images. A reasonable definition of the vehicle coordinate system in combination with the use of epipolar geometry reduces the complexity to parameterize the underlying homography matrix. The extrinsic parameters are determined analytically by two correctly matched feature points located on the road surface. The final parameter set is determined by a recursive filter which considers various estimates over time. Results with a real-world video sequence indicate that the method is comparable to classical offline calibration techniques using objects of known geometry.

In risk assessment applications well informed decisions are made based on huge amounts of multi-dimensional data. In many domains not only the risk of a wrong decision, but in particular the trade-off between the costs of possible decisions are of utmost importance. In this paper we describe a framework tightly integrating interactive visual exploration with machine learning to support the decision making process. The proposed approach uses a series of interactive 2D visualizations of numeric and ordinal data combined with visualization of classification models. These series of visual elements are further linked to the classifier's performance visualized using an interactive performance curve. An interactive decision point on the performance curve allows the decision maker to steer the classification model and instantly identify the critical, cost changing data elements, in the various linked visualizations. The critical data elements are represented as images in order to trigger associations related to the knowledge of the expert. In this context the data visualization and classification results are not only linked together, but are also linked back to the classification model. Such a visual analytics framework allows the user to interactively explore the costs of his decisions for different settings of the model and accordingly use the most suitable classification model and make more informed and reliable decisions. A case study on data from the Forensic Psychiatry domain reveals the usefulness of the suggested approach.

Many real-world problems involve time-oriented data. Time data is different from other kinds of data--explicitly harnessing the structures of time in visualizations can guide and support users’ visual analysis processes. State-of-the-art visualizations hardly take advantage of the structures of time to aid users in understanding and exploring the data. To bring more flexibility to the analysis process, we have developed interactive visual methods incorporating the structures of time within a pixel-based visualization called GROOVE (granular overview overlay). GROOVE uses different techniques to visualize time-oriented data by overlaying several time granularities in one visualization and provides interactive operators, which utilize the structures of time in different ways to capture and explore time-oriented data.

A test bed for research and teaching in fault-tolerant control (FTC) systems is presented. The laboratory plant is based on an industrial reverse osmosis desalination plant equipped with standardized components, which introduces more realism and robustness into the experiments. This paper describes the plant, the mathematical model of the system and an illustrative experiment. Moreover, the problem of choosing the hardware/software platform for the real-time operation is addresses.

Recently, Web browser based applications have become very popular in many domains. However, the specific requirements of interactive Information Visualization (InfoVis) applications in terms of graphics performance and interactivity have not yet been investigated systematically in this context. In order to assess browser-based application platforms, we provide a systematic comparison of server-based rendering, Java Applets, Flash, and Silverlight from several points of view. We aim to aid InfoVis developers in choosing the appropriate technology for their needs.

Empirical comparisons and categorizations of information visualization tools lack important considerations: the former undervalue the need for a theoretical background, and the latter tend to have too much distance from the user because they do not consider definite user tasks. Therefore, our work combines these approaches and presents the results of both a qualitative evaluation and a recently published categorization. We focus on the visualization of temporal data and reveal that current tools realize only a small part of the visualization possibilities in this field.

The study of isolated populations is a promising approach for the identification of genes conferring susceptibility to disease. Due to the complex genealogies of such populations epidemiological studies are challenging.We present ongoing research to using a visual analytics based approach for the identification of diseases that cluster in families, risk factors and heritability patterns.

Transforming a text-based clinical guideline in a computer-interpretable form is a time-consuming and demanding task due to the various users involved, who have different technical and medical background. In the past, different guideline representation languages and supporting tools have been developed, however, these approaches seldom address the various users' demands and needs in the different steps of the guidelines' life cycle. Our approach is oriented on the guideline life-cycle and takes the requirements and the interactions of the various actors into account to formalize guidelines in a computer-interpretable guideline representation by using a semi-automatic way based on NLP techniques. We analyzed the guideline life-cycle and the roles of the actors to build such a model. This model is prototypical implemented and showed the usefulness and utility for the various users.

Fault-tolerant controllers with online accommodation using model predictive control, linear quadratic controller and the pseudo inverse method are compared. In a real-time implementation, controllers are tested under actuator faults like saturation, freezing and total loss as well as under a structural fault. The advantages and disadvantages as well as the limits of the algorithms are considered.

The evaluation of Information Visualization (InfoVis) techniques can help to identify specific strengths and weaknesses of these methods. The following article describes the results of an empirical study assessing the contribution of an interactive InfoVis method based on a spring metaphor (GRAVI), exploratory data analysis (EDA) and machine learning (ML) to ease understanding. The application domain is the psychotherapeutic treatment of anorectic young women. The three methods are supposed to support the therapists in finding the variables which influence success or failure of the therapy. To conduct the evaluation we developed a report system which helped subjects to formulate and document in a self-directed manner the insights they gained when using the three methods. The results indicate that the three methods are complementary and should be used in conjunction.

Clinical practice guidelines are widely used to support medical staff in treatment planning and decision-making, whereas, the classification of different recommendations in the CPGs are one of the most important information sources to use. However, there is a lack of consensus amongst guideline developers, regarding those classification schemes. To address this problem, we mapped the different graded and ungraded evidence information used by different guideline developing organizations into our meta schema. In this paper we describe how guideline representation languages, such as Asbru and PROforma can be extended to model our meta schema.