This design study focuses on the analysis of a time sequence of categorical sequences. Such data is relevant for the geoscientific research field of landscape and climate development. It results from microscopic analysis of lake sediment cores. The goal is to gain hypotheses about landscape evolution and climate conditions in the past. To this end, geoscientists identify which categorical sequences are similar in the sense that they indicate similar conditions. Categorical sequences are similar if they have similar meaning (semantic similarity) and appear in similar time periods (temporal similarity). For data sets with many different categorical sequences, the task to identify similar sequences becomes a challenge. Our contribution is a tailored visual analysis concept that effectively supports the analytical process. Our visual interface comprises coupled visualizations of semantics and temporal context for the exploration and assessment of the similarity of categorical sequences. Integrated automatic methods reduce the analytical effort substantially. They (1) extract unique sequences in the data and (2) rank sequences by a similarity measure during the search for similar sequences. We evaluated our concept by demonstrations of our prototype to a larger audience and hands-on analysis sessions for two different lakes. According to geoscientists, our approach fills an important methodological gap in the application domain.

Abstract An established approach to studying interrelations between two non‐stationary time series is to compute the ‘windowed’ cross‐correlation (WCC). The time series are divided into intervals and the cross‐correlation between corresponding intervals is calculated. The outcome is a matrix that describes the correlation between two time series for different intervals and varying time lags. This important technique can only be used to compare two single time series. However, many applications require the comparison of ensembles of time series. Therefore, we propose a visual analytics approach that extends the WCC to support a correlation‐based comparison of two ensembles of time series. We compute the pairwise WCC between all time series from the two ensembles, which results in hundreds of thousands of WCC matrices. Statistical measures are used to derive a concise description of the time‐varying correlations between the ensembles as well as the uncertainty of the correlation values. We further introduce a visually scalable overview visualization of the computed correlation and uncertainty information. These components are combined with multiple linked views into a visual analytics system to support configuration of the WCC as well as detailed analysis of correlation patterns between two ensembles. Two use cases from very different domains, cognitive science and paleoclimatology, demonstrate the utility of our approach.

Researchers assess the quality of an ocean model by comparing its output to that of a previous model version or to observations. One objective of the comparison is to detect and to analyze differences and similarities between both data sets regarding geophysical processes, such as particular ocean currents. This task involves the analysis of thousands or hundreds of thousands of geographically referenced temporal profiles in the data. To cope with the amount of data, modelers combine aggregation of temporal profiles to single statistical values with visual comparison. Although this strategy is based on experience and a well-grounded body of expert knowledge, our discussions with domain experts have shown that it has two limitations: (1) using a single statistical measure results in a rather limited scope of the comparison and in significant loss of information, and (2) the decisions modelers have to make in the process may lead to important aspects being overlooked. In this article, we propose a Visual Analytics approach that broadens the scope of the analysis, reduces subjectivity, and facilitates comparison of the two data sets. It comprises three steps: First, it allows modelers to consider many aspects of the temporal behavior of geophysical processes by conducting multiple clusterings of the temporal profiles in each data set. Modelers can choose different features describing the temporal behavior of relevant processes, clustering algorithms, and parameterizations. Second, our approach consolidates the clusterings of one data set into a single clustering via a clustering ensembles approach. The consolidated clustering presents an overview of the geospatial distribution of temporal behavior in a data set. Third, a visual interface allows modelers to compare the two consolidated clusterings. It enables them to detect clusters of temporal profiles that represent geophysical processes and to analyze differences and similarities between two data sets. This work is the result of a close collaboration with ocean modelers. They employed our concept to find aspects of improvement in a new version of the Ocean Model for Circulation and Tides (OMCT).

Visualization techniques often use color to present categorical differences to a user. When selecting a color palette, the perceptual qualities of color need careful consideration. Large coherent groups visually suppress smaller groups and are often visually dominant in images. This paper introduces the concept of class visibility used to quantitatively measure the utility of a color palette to present coherent categorical structure to the user. We present a color optimization algorithm based on our class visibility metric to make categorical differences clearly visible to the user. We performed two user experiments on user preference and visual search to validate our visibility measure over a range of color palettes. The results indicate that visibility is a robust measure, and our color optimization can increase the effectiveness of categorical data visualizations.

We present a Visual Analytics approach that addresses the detection of interesting patterns in numerical time series, specifically from environmental sciences. Crucial for the detection of interesting temporal patterns are the time scale and the starting points one is looking at. Our approach makes no assumption about time scale and starting position of temporal patterns and consists of three main steps: an algorithm to compute statistical values for all possible time scales and starting positions of intervals, visual identification of potentially interesting patterns in a matrix visualization, and interactive exploration of detected patterns. We demonstrate the utility of this approach in two scientific scenarios and explain how it allowed scientists to gain new insight into the dynamics of environmental systems.

Abstract Many visualization techniques involve mapping high‐dimensional data spaces to lower‐dimensional views. Unfortunately, mapping a high‐dimensional data space into a scatterplot involves a loss of information; or, even worse, it can give a misleading picture of valuable structure in higher dimensions. In this paper, we propose class consistency as a measure of the quality of the mapping. Class consistency enforces the constraint that classes of n–D data are shown clearly in 2–D scatterplots. We propose two quantitative measures of class consistency, one based on the distance to the class's center of gravity, and another based on the entropies of the spatial distributions of classes. We performed an experiment where users choose good views, and show that class consistency has good precision and recall. We also evaluate both consistency measures over a range of data sets and show that these measures are efficient and robust.

The research reported in this paper focuses on integrating analytical and visual methods in order to explore complex patterns in geo‐related multivariate data sets and to understand the changes in patterns over time. The goal is to provide techniques that are able to analyse real‐world Data Warehouses, a typical architecture to manage such geo‐related multidimensional data sets, in order to support the analyst's decision‐making process. Challenges arise because real‐world applications usually have to deal with millions of records, with dozens of dimensions, and spatio‐temporal context. Therefore, a tight integration of automated analysis and interactive visualizations is needed (as proposed in the context of Visual Analytics). Our approach uses the well‐studied capabilities provided by Data Warehouses supporting knowledge discovery and decision‐making to analyse spatio‐temporal behaviour of pattern in high‐dimensional spaces. The topic of the paper is to show possible interplays between automated analysis and geo‐spatial visualization.

Companies trading stocks need to store information on stock prices over specific time intervals, which results in very large databases. Large quantities of numerical data (thousands of records) are virtually impossible to understand quickly and require the use of a visual model, since that is the fastest way for a human brain to absorb those enormous collections of data. However, little work has been done on verifying which visualizations are more suitable to represent these data sets. Such work is of crucial importance, since it enables us to identify those useful visual models and, in addition, opens our minds to new research possibilities. This paper presents an empirical study of different visualizations, that have been employed for stock market data, by comparing the results obtained by all studied techniques in typical exploratory data analysis tasks. This work provides several research contributions to the design of advanced visual data exploration interfaces.

The information revolution is creating and publishing vast data sets, such as records of business transactions, environmental statistics and census demographics. In many application domains, this data is collected and indexed by geo-spatial location. The discovery of interesting patterns in such databases through visual analytics is a key to turn this data into valuable information. Challenges arise because newly available geo-spatial data sets often have millions of records, or even far more, they are from multiple and heterogeneous data sources, and the output devices have significantly changed, e.g. high-resolution pixilated displays are increasingly available in both wall-sized and desktop units. New techniques are needed to cope with this scale. In this paper, we focus on ways to increase the scalability of pixel-based visual interfaces by adding task on hands scenarios that tightly integrate the data analyst into the exploration of geo-spatial data sets

In many applications, data is collected and indexed by geo-spatial location. Discovering interesting patterns through visualization is an important way of gaining insight about such data. A previously proposed approach is to apply local placement functions such as PixelMaps that transform the input data set into a solution set that preserves certain constraints while making interesting patterns more obvious and avoid data loss from overplotting. In experience, this family of spatial transformations can reveal fine structures in large point sets, but it is sometimes difficult to relate those structures to basic geographic features such as cities and regional boundaries. Recent information visualization research has addressed other types of transformation functions that make spatially-transformed maps with recognizable shapes. These types of spatial-transformation are called global shape functions. In particular, cartogram-based map distortion has been studied. On the other hand, cartogram-based distortion does not handle point sets readily. In this study, we present a framework that allows the user to specify a global shape function and a local placement function. We combine cartogram-based layout (global shape) with PixelMaps (local placement), obtaining some of the benefits of each toward improved exploration of dense geo-spatial data sets.

In today’s world, e-mail has become one of the most important means of communication in business and private lives due to its efficiency. However, the problems start as soon as mail volumes go beyond the scope of human information processing capabilities. Firstly, time does not allow for leaving certain messages unanswered for a long time, and in certain cases, for reading all messages. Secondly, the dilemma of electronic filters leaves a choice of too many junk mails getting through versus a risk of solicited mails being dumped. In this paper we present a new interactive visual data mining approach for analyzing individual e-mail communication. It combines classical visual analytics (help to identify pattern such as peaks and trends over time) with geo-spatial map distortions (help to understand the routes of e-mails). Experiments show that our visual e-mail explorer produces useful and interesting visualizations of large collections of e-mail and is practical for exploring temporal and geo-spatial patterns hidden in the e-mail data.

This paper introduces a new approach for visualizing multidimensional time-referenced data sets, called Circle View. The Circle View technique is a combination of hierarchical visualization techniques, such as treemaps [6], and circular layout techniques such as Pie Charts and Circle Segments [2]. The main goal is to compare continuous data changing their characteristics over time in order to identify patterns, exceptions and similarities in the data.To achieve this goal Circle View is a intuitive and easy to understand visualization interface to enable the user very fast to acquire the information needed. This is an important feature for fast changing visualization caused by time related data streams. Circle View supports the visualization of the changing characteristics over time, to allow the user the observation of changes in the data. Additionally it provides user interaction and drill down mechanism depending on user demands for a effective exploratory data analysis. There is also the capability of exploring correlations and exceptions in the data by using similarity and ordering algorithms.

PixelMaps are a new pixel-oriented visual data mining technique for large spatial datasets. They combine kernel-density-based clustering with pixel-oriented displays to emphasize clusters while avoiding overlap in locally dense point sets on maps. Because a full evaluation of density functions is prohibitively expensive, we also propose an efficient approximation, Fast-PixelMap, based on a synthesis of the quadtree and gridfile data structures.