Sebastian Mittelstädt

Special Section on Visual Analytics: Anomaly detection for visual analytics of power consumption data

Commercial buildings are significant consumers of electrical power. Also, energy expenses are an increasing cost factor. Many companies therefore want to save money and reduce their power usage. Building administrators have to first understand the power consumption behavior, before they can devise strategies to save energy. Second, sudden unexpected changes in power consumption may hint at device failures of critical technical infrastructure. The goal of our research is to enable the analyst to understand the power consumption behavior and to be aware of unexpected power consumption values. In this paper, we introduce a novel unsupervised anomaly detection algorithm and visualize the resulting anomaly scores to guide the analyst to important time points. Different possibilities for visualizing the power usage time series are presented, combined with a discussion of the design choices to encode the anomaly values. Our methods are applied to real-world time series of power consumption, logged in a hierarchical sensor network.

A visual analytics approach for peak-preserving prediction of large seasonal time series

Time series prediction methods are used on a daily basis by analysts for making important decisions. Most of these methods use some variant of moving averages to reduce the number of data points before prediction. However, to reach a good prediction in certain applications (e.g., power consumption time series in data centers) it is important to preserve peaks and their patterns. In this paper, we introduce automated peak‐preserving smoothing and prediction algorithms, enabling a reliable long term prediction for seasonal data, and combine them with an advanced visual interface: (1) using high resolution cell‐based time series to explore seasonal patterns, (2) adding new visual interaction techniques (multi‐scaling, slider, and brushing & linking) to incorporate human expert knowledge, and (3) providing both new visual accuracy color indicators for validating the predicted results and certainty bands communicating the uncertainty of the prediction. We have integrated these techniques into a well‐fitted solution to support the prediction process, and applied and evaluated the approach to predict both power consumption and server utilization in data centers with 70–80% accuracy.

Visual Analysis of Time-Series Similarities for Anomaly Detection in Sensor Networks

We present a system to analyze time‐series data in sensor networks. Our approach supports exploratory tasks for the comparison of univariate, geo‐referenced sensor data, in particular for anomaly detection. We split the recordings into fixed‐length patterns and show them in order to compare them over time and space using two linked views. Apart from geo‐based comparison across sensors we also support different temporal patterns to discover seasonal effects, anomalies and periodicities.

Methods for compensating contrast effects in information visualization

Color, as one of the most effective visual variables, is used in many techniques to encode and group data points according to different features. Relations between features and groups appear as visual patterns in the visualization. However, optical illusions may bias the perception at the first level of the analysis process. For instance, in pixel‐based visualizations contrast effects make pixels appear brighter if surrounded by a darker area, which distorts the encoded metric quantity of the data points. Even if we are aware of these perceptual issues, our visual cognition system is not able to compensate these effects accurately. To overcome this limitation, we present a color optimization algorithm based on perceptual metrics and color perception models to reduce physiological contrast or color effects. We evaluate our technique with a user study and find that the technique doubles the accuracy of users comparing and estimating color encoded data values. Since the presented technique can be used in any application without adaption to the visualization itself, we are able to demonstrate its effectiveness on data visualizations in different domains.

A survey and task-based quality assessment of static 2D colormaps

Color is one of the most important visual variables since it can be combined with any other visual mapping to encode information without using additional space on the display. Encoding one or two dimensions with color is widely explored and discussed in the field. Also mapping multi-dimensional data to color is applied in a vast number of applications, either to indicate similar, or to discriminate between different elements or (multi-dimensional) structures on the screen. A variety of 2D colormaps exists in literature, covering a large variance with respect to different perceptual aspects. Many of the colormaps have a different perspective on the underlying data structure as a consequence of the various analysis tasks that exist for multivariate data. Thus, a large design space for 2D colormaps exists which makes the development and use of 2D colormaps cumbersome. According to our literature research, 2D colormaps have not been subject of in-depth quality assessment. Therefore, we present a survey of static 2D colormaps as applied for information visualization and related fields. In addition, we map seven devised quality assessment measures for 2D colormaps to seven relevant tasks for multivariate data analysis. Finally, we present the quality assessment results of the 2D colormaps with respect to the seven analysis tasks, and contribute guidelines about which colormaps to select or create for each analysis task.

An Integrated In-Situ Approach to Impacts from Natural Disasters on Critical Infrastructures

Natural disasters can have a devastating effect on critical infrastructures, especially in case of cascading effects among multiple infrastructures such as the electric power grid, the communication network, and the road network. While there exist detailed models for individual types of infrastructures such as electric power grids, these do not encompass the various interconnections and interdependencies to other networks. Cascading effects are hard to discover and often the root causes of problems remain unclear. In order to enable real-time situational awareness for operational risk management one needs to be aware of the broader context of events. In this paper, we present a unique visual analytics control room system that integrates the separate visualizations of the different network infrastructures with social media analysis and mobile in-situ analysis to help to monitor the critical infrastructures, detecting cascading effects, performing root cause analyses, and managing the crisis response. Both the social media analysis and the mobile in-situ analysis are important components for an effective understanding of the crisis and an efficient crisis response. Our system provides a mechanism for conjoining the available information of different infrastructures and social media as well as mobile in-situ analysis in order to provide unified views and analytical tools for monitoring, planning, and decision support. A realistic use case scenario based on real critical infrastructures as well as our qualitative study with crisis managers shows the potential of our approach.

Efficient contrast effect compensation with personalized perception models

Color is one of the most effective visual variables and is frequently used to encode metric quantities. Contrast effects are considered harmful in data visualizations since they significantly bias our perception of colors. For instance, a gray patch appears brighter on a black background than on a white background. Accordingly, the perception of color‐encoded data items depends on the surround in the rendered visualization. A method that compensates for contrast effects has been presented previously, which significantly improves the users’ accuracy in reading and comparing color encoded data. The method utilizes established perception models to compensate for contrast effects, assuming an average human observer. In this paper, we provide experiments that show a significant difference in the perception of users. We introduce methods to personalize contrast effect compensation and show that this outperforms the original method with a user study. We, further, overcome the major limitation of the original method, which is a runtime of several minutes. With the use of efficient optimization and surrogate models, we are able to reduce runtime to milliseconds, making the method applicable in interactive visualizations.

Revisiting Perceptually Optimized Color Mapping for High-Dimensional Data Analysis

Colors is one of the most effective visual variables since it can be combined with other mappings and encode information without using any additional space on the display. An important example where expressing additional visual dimensions is direly needed is the analysis of high-dimensional data. The property of perceptual linearity is desirable in this application, because the user intuitively perceives clusters and relations among multi-dimensional data points. Many approaches use two-dimensional colormaps in their analysis, which are typically created by interpolating in RGB, HSV or CIELAB color spaces. These approaches share the problem that the resulting colors are either saturated and discriminative but not perceptual linear or vice versa. A solution that combines both advantages has been previously introduced by Kaski et al.; yet, this method is to date underutilized in Information Visualization according to our literature analysis. The method maps high-dimensional data points into the CIELAB color space by maintaining the relative perceived distances of data points and color discrimination. In this paper, we generalize and extend the method of Kaski et al. to provide perceptual uniform color mapping for visual analysis of high-dimensional data. Further, we evaluate the method and provide guidelines for different analysis tasks.

Bridging the Gap of Domain and Visualization Experts with a Liaison

We introduce the role Liaison for design study projects. With considerable expertise in visualization and the application domain, a Liaison can help to foster richer and more effective interdisciplinary communication in problem characterization, design, and evaluation processes. We characterize this role, provide a list of tasks of Liaison and visualization experts, and discuss concrete benefits and potential limitations based on our experience from multiple design studies. To illustrate our contributions we use as an example a molecular biology design study.

ColorCAT: Guided Design of Colormaps for Combined Analysis Tasks

Colormap design is challenging because the encoding must match the requirements of data and analysis tasks as well as the perception of the target user. A number of well-known tools exist to support the design of colormaps. ColorBrewer [HB03], for example, is a great resource to select colors for qualitative, sequential, and diverging data. PRAVDAColor [BRT95] and Tominski et al. [TFS08], for example, provide valuable guidelines for single analysis tasks such as localization, identification, and comparison. However, for solving real world problems in most practical applications, single elementary analysis tasks are not sufficient but need to be combined. In this paper, we propose a methodology and tool to design colormaps for combined analysis tasks. We define color mapping requirements and develop a set of design guidelines. The visualization expert is integrated in the design process to incorporate his/her design requirements, which may depend on the application, culture, and aesthetics. Our ColorCAT tool guides novice and expert designers through the creation of colormaps and allows the exploration of the design space of color mapping for combined analysis tasks.