Dashboards are one of the most common use cases for data visualization, and their design and contexts of use are considerably different from exploratory visualization tools. In this paper, we look at the broad scope of how dashboards are used in practice through an analysis of dashboard examples and documentation about their use. We systematically review the literature surrounding dashboard use, construct a design space for dashboards, and identify major dashboard types. We characterize dashboards by their design goals, levels of interaction, and the practices around them. Our framework and literature review suggest a number of fruitful research directions to better support dashboard design, implementation, and use.

Analysts need interactive speed for exploratory analysis, but big data systems are often slow. With sampling, data systems can produce approximate answers fast enough for exploratory visualization, at the cost of accuracy and trust. We propose optimistic visualization, which approaches these issues from a user experience perspective. This method lets analysts explore approximate results interactively, and provides a way to detect and recover from errors later. Pangloss implements these ideas. We discuss design issues raised by optimistic visualization systems. We test this concept with five expert visualizers in a laboratory study and three case studies at Microsoft. Analysts reported that they felt more confident in their results, and used optimistic visualization to check that their preliminary results were correct.

A common workflow for visualization designers begins with a generative tool, like D3 or Processing, to create the initial visualization; and proceeds to a drawing tool, like Adobe Illustrator or Inkscape, for editing and cleaning. Unfortunately, this is typically a one-way process: once a visualization is exported from the generative tool into a drawing tool, it is difficult to make further, data-driven changes. In this paper, we propose a bridge model to allow designers to bring their work back from the drawing tool to re-edit in the generative tool. Our key insight is to recast this iteration challenge as a merge problem - similar to when two people are editing a document and changes between them need to reconciled. We also present a specific instantiation of this model, a tool called Hanpuku, which bridges between D3 scripts and Illustrator. We show several examples of visualizations that are iteratively created using Hanpuku in order to illustrate the flexibility of the approach. We further describe several hypothetical tools that bridge between other visualization tools to emphasize the generality of the model.

As real-time monitoring and analysis become increasingly important, researchers and developers turn to data stream management systems (DSMS's) for fast, efficient ways to pose temporal queries over their datasets. However, these systems are inherently complex, and even database experts find it difficult to understand the behavior of DSMS queries. To help analysts better understand these temporal queries, we developed StreamTrace, an interactive visualization tool that breaks down how a temporal query processes a given dataset, step-by-step. The design of StreamTrace is based on input from expert DSMS users; we evaluated the system with a lab study of programmers who were new to streaming queries. Results from the study demonstrate that StreamTrace can help users to verify that queries behave as expected and to isolate the regions of a query that may be causing unexpected results.

Many different domains collect event sequence data and rely on finding and analyzing patterns within it to gain meaningful insights. Current systems that support such queries either provide limited expressiveness, hinder exploratory workflows or present interaction and visualization models which do not scale well to large and multi-faceted data sets. In this paper we present (s|qu)eries (pronounced "Squeries"), a visual query interface for creating queries on sequences (series) of data, based on regular expressions. (s|qu)eries is a touch-based system that exposes the full expressive power of regular expressions in an approachable way and interleaves query specification with result visualizations. Being able to visually investigate the results of different query-parts supports debugging and encourages iterative query-building as well as exploratory work-flows. We validate our design and implementation through a set of informal interviews with data scientists that analyze event sequences on a daily basis.

Data scientists often conduct exploratory data analysis in scripting environments with a read-eval-print loop (REPL), like R, IPython or MATLAB. This user experience requires diligent management of execution and generates lengthy histories of unwanted command responses. This paper explores the alternative of live programming, a user experience in which the user's edits immediately and automatically update the script results-a “ripple” effect familiar from spreadsheets. Which user experience provides better support for exploratory data analysis, REPL or ripple? We conducted a controlled lab study with 15 data-experienced professionals. Each participant explored four datasets, two in each experience. The REPL sessions left histories with both significantly more data results and significantly more errors than the live sessions. However, both experiences produced comparable numbers of data results that participants self-rated as insightful. Participants largely preferred the live experience for its responsiveness and ability to keep the script content clean, but missed the visible history that a REPL provides.

Data scientists are increasingly working with live streaming data, for example, business telemetry and signals from wearable devices and the Internet of Things. Unfortunately, current tools for exploratory data analysis provide poor support for streaming data. This paper presents Tempe, a data science environment for temporal and streaming data. Tempe's extensible scripting environment allows for live programming, displays interactive, continually updating visualizations, and provides a uniform query language for both stored and live data. We discuss the streaming features of Tempe and evaluate our design choices with a deployment study at Microsoft with a product team who used Tempe continuously for six months.

In recent years many popular data visualizations have emerged that are created largely by designers whose main area of expertise is not computer science. Designers generate these visualizations using a handful of design tools and environments. To better inform the development of tools intended for designers working with data, we set out to understand designers' challenges and perspectives. We interviewed professional designers, conducted observations of designers working with data in the lab, and observed designers working with data in team settings in the wild. A set of patterns emerged from these observations from which we extract a number of themes that provide a new perspective on design considerations for visualization tool creators, as well as on known engineering problems.

We often use datasets that reflect samples, but many visualization tools treat data as full populations. Uncertain visualizations are good at representing data distributions emerging from samples, but are more limited in allowing users to carry out decision tasks. This is because tasks that are simple on a traditional chart (e.g. "compare two bars") become a complex probabilistic task on a chart with uncertainty. We present guidelines for creating visual annotations for solving tasks with uncertainty, and an implementation that addresses five core tasks on a bar chart. A preliminary user study shows promising results: that users have a justified confidence in their answers with our system.

Labeling data is a seemingly simple task required for training many machine learning systems, but is actually fraught with problems. This paper introduces the notion of concept evolution, the changing nature of a person's underlying concept (the abstract notion of the target class a person is labeling for, e.g., spam email, travel related web pages) which can result in inconsistent labels and thus be detrimental to machine learning. We introduce two structured labeling solutions, a novel technique we propose for helping people define and refine their concept in a consistent manner as they label. Through a series of five experiments, including a controlled lab study, we illustrate the impact and dynamics of concept evolution in practice and show that structured labeling helps people label more consistently in the presence of concept evolution than traditional labeling.

As more applications move from the desktop to touch devices like tablets, designers must wrestle with the costs of porting a design with as little revision of the UI as possible from one device to the other, or of optimizing the interaction per device. We consider the tradeoffs between two versions of a UI for working with data on a touch tablet. One interface is based on using the conventional desktop metaphor (WIMP) with a control panel, push buttons, and checkboxes -- where the mouse click is effectively replaced by a finger tap. The other interface (which we call FLUID) eliminates the control panel and focuses touch actions on the data visualization itself. We describe our design process and evaluation of each interface. We discuss the significantly better task performance and preference for the FLUID interface, in particular how touch design may challenge certain assumptions about the performance benefits of WIMP interfaces that do not hold on touch devices, such as the superiority of gestural vs. control panel based interaction.

Exploratory analysis on big data requires us to rethink data management across the entire stack -- from the underlying data processing techniques to the user experience. We demonstrate Stat! -- a visualization and analytics environment that allows users to rapidly experiment with exploratory queries over big data. Data scientists can use Stat! to quickly refine to the correct query, while getting immediate feedback after processing a fraction of the data. Stat! can work with multiple processing engines in the backend; in this demo, we use Stat! with the Microsoft StreamInsight streaming engine. StreamInsight is used to generate incremental early results to queries and refine these results as more data is processed. Stat! allows data scientists to explore data, dynamically compose multiple queries to generate streams of partial results, and display partial results in both textual and visual form.

Conveying a narrative with visualizations often requires choosing an order in which to present visualizations. While evidence exists that narrative sequencing in traditional stories can affect comprehension and memory, little is known about how sequencing choices affect narrative visualization. We consider the forms and reactions to sequencing in narrative visualization presentations to provide a deeper understanding with a focus on linear, 'slideshow-style' presentations. We conduct a qualitative analysis of 42 professional narrative visualizations to gain empirical knowledge on the forms that structure and sequence take. Based on the results of this study we propose a graph-driven approach for automatically identifying effective sequences in a set of visualizations to be presented linearly. Our approach identifies possible transitions in a visualization set and prioritizes local (visualization-to-visualization) transitions based on an objective function that minimizes the cost of transitions from the audience perspective. We conduct two studies to validate this function. We also expand the approach with additional knowledge of user preferences for different types of local transitions and the effects of global sequencing strategies on memory, preference, and comprehension. Our results include a relative ranking of types of visualization transitions by the audience perspective and support for memory and subjective rating benefits of visualization sequences that use parallelism as a structural device. We discuss how these insights can guide the design of narrative visualization and systems that support optimization of visualization sequence.

Temporal events with multiple sets of metadata attributes, i. e., facets, are ubiquitous across different domains. The capabilities of efficiently viewing and comparing events data from various perspectives are critical for revealing relationships, making hypotheses, and discovering patterns. In this paper, we present TimeSlice, an interactive faceted visualization of temporal events, which allows users to easily compare and explore timelines with different attributes on a set of facets. By directly manipulating the filtering tree, a dynamic visual representation of queries and filters in the facet space, users can simultaneously browse the focused timelines and their contexts at different levels of detail, which supports efficient navigation of multi-dimensional events data. Also presented is an initial evaluation of TimeSlice with two datasets - famous deceased people and US daily flight delays.

Queries over large scale (petabyte) data bases often mean waiting overnight for a result to come back. Scale costs time. Such time also means that potential avenues of exploration are ignored because the costs are perceived to be too high to run or even propose them. With sampleAction we have explored whether interaction techniques to present query results running over only incremental samples can be presented as sufficiently trustworthy for analysts both to make closer to real time decisions about their queries and to be more exploratory in their questions of the data. Our work with three teams of analysts suggests that we can indeed accelerate and open up the query process with such incremental visualizations.

Co-located collaboration can be extremely valuable during complex visual analytics tasks. We present an exploratory study of a system designed to support collaborative visual analysis tasks on a digital tabletop display. Fifteen participant pairs employed Cambiera, a visual analytics system, to solve a problem involving 240 digital documents. Our analysis, supported by observations, system logs, questionnaires, and interview data, explores how pairs approached the problem around the table. We contribute a unique, rich understanding of how users worked together around the table and identify eight types of collaboration styles that can be used to identify how closely people work together while problem solving. We show how the closeness of teams’ collaboration and communication influenced how they performed on the task overall. We further discuss the role of the tabletop for visual analytics tasks and derive design implications for future co-located collaborative tabletop problem solving systems.

Cambiera is a tabletop system designed for co-located collaborative visual analytics. As a tabletop system, Cambiera encourages analysts to face each other around the tabletop to analyze large text document collections collaboratively. Cambiera allows analysts to search for documents and read them, organize documents on the tabletop, and to monitor each other's' work. The video illustrates the major collaborative features of Cambiera.

The goal of a research institution is, ultimately, to share and disseminate knowledge. Yet the sheer volume of information produced by large institutions makes it challenging to keep track of the vast knowledge within. Information on who knows what is often scattered across multiple sources and media. Expertise tracking systems allow users to search for people who know answers, but do not support serendipitous discovery. To help visitors and researchers alike develop awareness of research activities, we have designed ResearchWave---a large-display ambient visualization, installed in the social spaces of a research institution. ResearchWave represents information on research activities in a lightweight and aesthetically pleasing manner. Research-Wave is based on a "walk up and use" approach: it uses multiple levels of visual encodings to engage people while allowing them to learn more with each novel encounter. In this paper, we report our design process, first prototype and lessons learned from initial user feedback.

Interactive clustering refers to situations in which a human labeler is willing to assist a learning algorithm in automatically clustering items. We present a related but somewhat different task, assisted clustering, in which a user creates explicit groups of items from a large set and wants suggestions on what items to add to each group. While the traditional approach to interactive clustering has been to use metric learning to induce a distance metric, our situation seems equally amenable to classification. Using clusterings of documents from human subjects, we found that one or the other method proved to be superior for a given cluster, but not uniformly so. We thus developed a hybrid mechanism for combining the metric learner and the classifier. We present results from a large number of trials based on human clusterings, in which we show that our combination scheme matches and often exceeds the performance of a method which exclusively uses either type of learner.

In order to use new visualizations, most toolkits require application developers to rebuild their applications and distribute new versions to users. The WebCharts Framework take a different approach by hosting JavaScript from within an application and providing a standard data and events interchange. In this way, applications can be extended dynamically, with a wide variety of visualizations. We discuss the benefits of this architectural approach, contrast it to existing techniques, and give a variety of examples and extensions of the basic system.

People regularly interact with different representations of Web pages. A person looking for new information may initially find a Web page represented as a short snippet rendered by a search engine. When he wants to return to the same page the next day, the page may instead be represented by a link in his browser history. Previous research has explored how to best represent Web pages in support of specific task types, but, as we find in this paper, consistency in representation across tasks is also important. We explore how different representations are used in a variety of contexts and present a compact representation that supports both the identification of new, relevant Web pages and the re-finding of previously viewed pages.

Abstract Many real‐world analysis tasks can benefit from the combined efforts of a group of people. Past research has shown that to design visualizations for collaborative visual analytics tasks, we need to support both individual as well as joint analysis activities. We present Cambiera, a tabletop visual analytics tool that supports individual and collaborative information foraging activities in large text document collections. We define collaborative brushing and linking as an awareness mechanism that enables analysts to follow their own hypotheses during collaborative sessions while still remaining aware of the group's activities. With Cambiera, users are able to collaboratively search through documents, maintaining awareness of each others' work and building on each others' findings.

The research presented in this paper compares user-generated and automatic graph layouts. Following the methods suggested by van Ham et al. (2008), a group of users generated graph layouts using both multi-touch interaction on a tabletop display and mouse interaction on a desktop computer. Users were asked to optimize their layout for aesthetics and analytical tasks with a social network. We discuss characteristics of the user-generated layouts and interaction methods employed by users in this process. We then report on a web-based study to compare these layouts with the output of popular automatic layout algorithms. Our results demonstrate that the best of the user-generated layouts performed as well as or better than the physics-based layout. Orthogonal and circular automatic layouts were found to be considerably less effective than either the physics-based layout or the best of the user-generated layouts. We highlight several attributes of the various layouts that led to high accuracy and improved task completion time, as well as aspects in which traditional automatic layout methods were unsuccessful for our tasks.

The recent introduction of simple, web-based geographic visualization interfaces has unleashed a tidal wave of new geographic content now available on the Internet. There has been enormous attention on the development of data interchange standards and programming interfaces that make all this content interoperable, but far less thought about how the user experience should change when users have their choice of 10,000 maps.To inform the design of online mapping systems, we investigate the case of queries that require correlation of multiple maps---that is, discovery and synthesis of several map layers. We based our study on interviews with expert users of maps: archivists and librarians. This paper describes our user-task taxonomy distilled from these interviews, and presents MapSynthesizer, a prototype system that allows users to efficiently query, discover, and integrate many maps from a corpus of thousands.

Animation has been used to show trends in multi-dimensional data. This technique has recently gained new prominence for presentations, most notably with Gapminder Trendalyzer. In Trendalyzer, animation together with interesting data and an engaging presenter helps the audience understand the results of an analysis of the data. It is less clear whether trend animation is effective for analysis. This paper proposes two alternative trend visualizations that use static depictions of trends: one which shows traces of all trends overlaid simultaneously in one display and a second that uses a small multiples display to show the trend traces side-by-side. The paper evaluates the three visualizations for both analysis and presentation. Results indicate that trend animation can be challenging to use even for presentations; while it is the fastest technique for presentation and participants find it enjoyable and exciting, it does lead to many participant errors. Animation is the least effective form for analysis; both static depictions of trends are significantly faster than animation, and the small multiples display is more accurate.

Understanding how people use online maps allows data acquisition teams to concentrate their efforts on the portions of the map that are most seen by users. Online maps represent vast databases, and so it is insufficient to simply look at a list of the most-accessed URLs. Hotmap takes advantage of the design of a mapping system's imagery pyramid to superpose a heatmap of the log files over the original maps. Users' behavior within the system can be observed and interpreted. This paper discusses the imagery acquisition task that motivated Hotmap, and presents several examples of information that Hotmap makes visible. we discuss the design choices behind Hotmap, including logarithmic color schemes; low-saturation background images; and tuning images to explore both infrequently-viewed and frequently-viewed spaces.

Ten years ago, Whittaker and Sidner [8] published research on email overload, coining a term that would drive a research area that continues today. We examine a sample of 600 mailboxes collected at a high-tech company to compare how users organize their email now to 1996. While inboxes are roughly the same size as in 1996, our population's email archives have grown tenfold. We see little evidence of distinct strategies for handling email; most of our users fall into a middle ground. There remains a need for future innovations to help people manage growing archives of email and large inboxes.

Everyday work frequently involves coordinating and collaborating with others, but the structure of collaboration is largely invisible to conventional desktop applications. We are exploring ways to support everyday collaboration by allowing applications access to the social, organizational, and temporal settings within which work is conducted. In this paper, we present two generations of systems supporting everyday collaboration, focusing on ways to recover and represent the temporal and social structures of online activity.

We present a Java-based framework, SWAMI (Shared Wisdom through the Amalgamation of Many Interpretations) for building and studying collaborative filtering systems. SWAMI consists of three components: a prediction engine, an evaluation system, and a visualization component. The prediction engine provides a common interface for implementing different prediction algorithms. The evaluation system provides a standardized testing methodology and metrics for analyzing the accuracy and run-time performance of prediction algorithms. The visualization component suggests how graphical representations can inform the development and analysis of prediction algorithms. We demonstrate SWAMI on the Each Movie data set by comparing three prediction algorithms: a traditional Pearson correlation-based method, support vector machines, and a new accurate and scalable correlation-based method based on clustering techniques.