Anthony C. Robinson

Visualizing Geospatial Information Uncertainty: What We Know and What We Need to Know

Developing reliable methods for representing and managing information uncertainty remains a persistent and relevant challenge to GIScience. Information uncertainty is an intricate idea, and recent examinations of this concept have generated many perspectives on its representation and visualization, with perspectives emerging from a wide range of disciplines and application contexts. In this paper, we review and assess progress toward visual tools and methods to help analysts manage and understand information uncertainty. Specifically, we report on efforts to conceptualize uncertainty, decision making with uncertainty, frameworks for representing uncertainty, visual representation and user control of displays of information uncertainty, and evaluative efforts to assess the use and usability of visual displays of uncertainty. We conclude by identifying seven key research challenges in visualizing information uncertainty, particularly as it applies to decision making and analysis.

SensePlace2: GeoTwitter analytics support for situational awareness

Geographically-grounded situational awareness (SA) is critical to crisis management and is essential in many other decision making domains that range from infectious disease monitoring, through regional planning, to political campaigning. Social media are becoming an important information input to support situational assessment (to produce awareness) in all domains. Here, we present a geovisual analytics approach to supporting SA for crisis events using one source of social media, Twitter. Specifically, we focus on leveraging explicit and implicit geographic information for tweets, on developing place-time-theme indexing schemes that support overview+detail methods and that scale analytical capabilities to relatively large tweet volumes, and on providing visual interface methods to enable understanding of place, time, and theme components of evolving situations. Our approach is user-centered, using scenario-based design methods that include formal scenarios to guide design and validate implementation as well as a systematic claims analysis to justify design choices and provide a framework for future testing. The work is informed by a structured survey of practitioners and the end product of Phase-I development is demonstrated / validated through implementation in SensePlace2, a map-based, web application initially focused on tweets but extensible to other media.

Combining Usability Techniques to Design Geovisualization Tools for Epidemiology

Designing usable geovisualization tools is an emerging problem in GIScience software development. We are often satisfied that a new method provides an innovative window on our data, but functionality alone is insufficient assurance that a tool is applicable to a problem in situ. As extensions of the static methods they evolved from, geovisualization tools are bound to enable new knowledge creation. We have yet to learn how to adapt techniques from interaction designers and usability experts toward our tools in order to maximize this ability. This is especially challenging because there is limited existing guidance for the design of usable geovisualization tools. Their design requires knowledge about the context of work within which they will be used, and should involve user input at all stages, as is the practice in any human-centered design effort. Toward that goal, we have employed a wide range of techniques in the design of ESTAT, an exploratory geovisualization toolkit for epidemiology. These techniques include; verbal protocol analysis, card-sorting, focus groups, and an in-depth case study. This paper reports the design process and evaluation results from our experience with the ESTAT toolkit.

Collaborative synthesis of visual analytic results

Visual analytic tools allow analysts to generate large collections of useful analytical results. We anticipate that analysts in most real world situations will draw from these collections when working together to solve complicated problems. This indicates a need to understand how users synthesize multiple collections of results. This paper reports the results of collaborative synthesis experiments conducted with expert geographers and disease biologists. Ten participants were worked in pairs to complete a simulated real-world synthesis task using artifacts printed on cards on a large, paper-covered workspace. Experiment results indicate that groups use a number of different approaches to collaborative synthesis, and that they employ a variety of organizational metaphors to structure their information. It is further evident that establishing common ground and role assignment are critical aspects of collaborative synthesis. We conclude with a set of general design guidelines for collaborative synthesis support tools.

The geoviz toolkit: using component-oriented coordination methods for geographic visualization and analysis

In this paper, we present the GeoViz Toolkit (GVT), an open-source, Internet-delivered program for geographic visualization and analysis that features a diverse set of software components which can be flexibly combined by users who do not have programming expertise. The design and architecture of the GVT allows us to address three key research challenges in geovisualization: allowing end users to create their own geovisualization and analysis component set on the fly, integrating geovisualization methods with spatial analysis methods, and making geovisualization applications sharable between users. Each of these tasks necessitates a robust yet flexible approach to intertool coordination. The coordination strategy developed for the GVT, called Introspective Observer Coordination, leverages and combines key advances in software engineering from the last decade, such as automatic introspection of objects, software design patterns, and reflective invocation of methods.

Leveraging Geospatially-Oriented Social Media Communications in Disaster Response

Geospatially-oriented social media communications have emerged as a common information resource to support crisis management. Our research compares the capabilities of two popular systems used to collect and visualize such information Project Epic’s Tweak the Tweet (TtT) and Ushahidi. Our research uses geospatially-oriented social media gathered by both projects during recent disasters to compare and contrast the frequency, content, and location components of contributed information to both systems. We compare how data was gathered and filtered, how spatial information was extracted and mapped, and the mechanisms by which the resulting synthesized information was shared with response and recovery organizations. In addition, we categorize the degree to which each platform in each disaster led to actions by first responders and emergency managers. Based on the results of our comparisons we identify key design considerations for future social media mapping tools to support crisis management.

Distributed usability evaluation of the Pennsylvania Cancer Atlas

BackgroundThe Pennsylvania Cancer Atlas (PA-CA) is an interactive online atlas to help policy-makers, program managers, and epidemiologists with tasks related to cancer prevention and control. The PA-CA includes maps, graphs, tables, that are dynamically linked to support data exploration and decision-making with spatio-temporal cancer data. Our Atlas development process follows a user-centered design approach. To assess the usability of the initial versions of the PA-CA, we developed and applied a novel strategy for soliciting user feedback through multiple distributed focus groups and surveys. Our process of acquiring user feedback leverages an online web application (e-Delphi). In this paper we describe the PA-CA, detail how we have adapted e-Delphi web application to support usability and utility evaluation of the PA-CA, and present the results of our evaluation.ResultsWe report results from four sets of users. Each group provided structured individual and group assessments of the PA-CA as well as input on the kinds of users and applications for which it is best suited. Overall reactions to the PA-CA are quite positive. Participants did, however, provide a range of useful suggestions. Key suggestions focused on improving interaction functions, enhancing methods of temporal analysis, addressing data issues, and providing additional data displays and help functions. These suggestions were incorporated in each design and implementation iteration for the PA-CA and used to inform a set of web-atlas design principles.ConclusionFor the Atlas, we find that a design that utilizes linked map, graph, and table views is understandable to and perceived to be useful by the target audience of cancer prevention and control professionals. However, it is clear that considerable variation in experience using maps and graphics exists and for those with less experience, integrated tutorials and help features are needed. In relation to our usability assessment strategy, we find that our distributed, web-based method for soliciting user input is generally effective. Advantages include the ability to gather information from users distributed in time and space and the relative anonymity of the participants while disadvantages include less control over when and how often participants provide input and challenges for obtaining rich input.

Visual Analysis of Historic Hotel Visitation Patterns

Understanding the space and time characteristics of human interaction in complex social networks is a critical component of visual tools for intelligence analysis, consumer behavior analysis, and human geography. Visual identification and comparison of patterns of recurring events is an essential feature of such tools. In this paper, we describe a tool for exploring hotel visitation patterns in and around Rebersburg, Pennsylvania from 1898-1900. The tool uses a wrapping spreadsheet technique, called reruns, to display cyclic patterns of geographic events in multiple overlapping natural and artificial calendars. Implemented as an improvise visualization, the tool is in active development through a iterative process of data collection, hypothesis, design, discovery, and evaluation in close collaboration with historical geographers. Several discoveries have inspired ongoing data collection and plans to expand exploration to include historic weather records and railroad schedules. Distributed online evaluations of usability and usefulness have resulted in numerous feature and design recommendations

Highlighting in Geovisualization

Coordinated view geovisualizations allow users to interactively pick and attend to data observations across multiple views. This is frequently supported by the transient application of a visual effect to an observation during a mouse selection or rollover. This technique, known as highlighting, is typically implemented using a dedicated bright and saturated color to outline observations. In this paper we present a range of possibilities for alternative approaches to color highlighting, beginning with examples from the range of available visual variables and moving beyond those options to other, non-visual variable methods such as the use of lines to connect highlighted observations. We also describe design criteria for highlighting methods that can be used to predict and test the suitability of different approaches, and apply those criteria to our set of proposed methods to identify potential good candidates for implementation in future systems. Next, we present a set of highlighting types that define basic ways in which highlighting methods can be combined, modified, and driven directly by data values. We conclude by outlining several broad research challenges for future work on the development and evaluation of highlighting methods in geographic visualization.

Maps and the geospatial revolution: teaching a massive open online course (MOOC) in geography

The massive open online course (MOOC) is a new approach for teaching online. MOOCs stand apart from traditional online classes in that they support thousands of learners through content and assessment mechanisms that can scale. A reason for their size is that MOOCs are free for anyone to take. Here we describe the design, development, and teaching of a MOOC called Maps and the Geospatial Revolution. We explore the geography of the student population, the experience of teaching a MOOC, and evaluate its impacts on learning. We conclude with several key challenges and opportunities we see for MOOCs in geography.