The A-PR Hypothesis (Autonomy and Pattern Recognition) suggests that knowledge
visualization is a critical tool to support the interpretation of information in collaborative
intelligence systems for cross-disciplinary problem-solving, learning and innovation.

click image for book info

One of the most important corporate initiatives to build collaborative intelligence through shared visualizations is the IBM Many Eyes initiative. Users build visualizations and share them with the community. Users rate visualizations; those deemed most useful rise to the top of the ratings.

On March 1, 2011 J Plunkett uploaded his visualization of total worldwide oil consumption by country (in barrels per day). Communities of interest grow around shared topics of discussion. One highly rated visualization (below) showed subprime loans at the height of the mortgage crisis.

The most effective visualizations go beyond representing information to uncover a dynamic and thought-provoking way to interact with that information so that learning and new insight result.

Many Eyes is identifying communities of practice who could take the next step to develop problem-solving innovation networks, using their visualization tools to support the practice of collaborative intelligence.

From Mapping Space to Mapping Concepts
The transition from pure geographical mapping to mapping concepts evolved, much as nature evolved, producing intermediary examples, hybrids that combined geographical representation with the representation of ideas. Below are three examples of hybrids:
• Minard's famous map of Napoleon's march to Moscow (called by Edward Tufte the best visualization ever made), which combines a geographical representation of the path of the march with information about temperature and the size of the surviving army;
• NASA visualization of the world, color-coded to show amount of leaf coverage;
• the renowned London subway map, which shows passengers only what they need to know: where train lines intersect so that changes can be made.

Minnard's map (1869) was bridge from the tradition of spatial mapping to concept mapping; it unites the two, telling the story of loss and suffering in a vivid representation of data. The width of the line represents troop strength (from 400,000 at the beginning to a mere 10,000 at the end). The chart at the bottom shows temperatures (all below freezing). The image combines a spatial representation with a data representation, showing how many men Napoleon had at the beginning, and at different stages during his campaign, and what the temperatures were.

The NASA Visualization of leaf density shows amounts of vegetation, which could be animated to show changes over the time.


The London subway map us a spatial abstraction focusing on what the passenger needs to find his way on the underground — the stations where train lines intersect.

Collaborative Intelligence in Ecosystem Forecasting
Ecosystem forecasting is supported by information visualization, e.g.

1. Visualization of Data, Indicators, and Thresholds
2. Collaborative Problem-Solving — Process Visualization & Management
3. Navigation and Search — User Interface & Knowledge Management Frameworks
4. Geospatial Visualization — Spatio-Temporal Representations

Visualization of Data, Indicators, and Thresholds
Outstanding visualization is the key to understanding how components interact in a complex system. Tim Nyerges reviews the challenge of visualizing sustainability in his paper: “Linked Visualizations in Sustainability Modeling: An Approach Using Participatory GIS for Decision Support.” 

Three visualizations representing sustainability issues: 

1. Concept mapping. The diagram below, used by the USGS Decision Support Systems (DSS) team for Tahoe demonstrates how visualization can show complex relationships. The USGS DSS team has used concept maps to analyze a range of factors affecting the Tahoe Basin (Halsing, Hessenflow, and Wein 2004), associated with the Tahoe Constrained Optimization Model (TCOM) (Bernknopf et al. 2003). 

Example of visual conceptual models developed for indicator analysis. 

In the directed graph above, nodes represent:

• target category indicators,
• subordinate indicators affected by the primary indicators,
• control indicators, and
• threshold indicators.

Above, Ramana Rao “Environmental Issues” from Richard Saul Wurman in Understanding USA
effectively combines icons, with photgraphic imagery with explanatory text to show a land use cycle.

For visualizing hierachical data:

Danny Holten (2006), Hierarchical Edge Bundles: Visualization of Adjacency Relations in Hierarchical Data,
IEEE Transactions on Visualization and Computer Graphics, vol. 12, no 5, pp. 741-748. implemented in VTK 

Within massive, dynamically changing information spaces, visual analytics research concentrates on support for perceptual and cognitive capacity to support autonomous users to recognize patterns, and possibly discover the unexpected in complex information spaces. Technologies resulting from visual analytics can be applied in a range of fields, with particular current emphasis on biology and national security. Visual representations of complexity allow us to use our sophisticated pattern recognition capacities to find new meaning in data, as in this graphic by Michael Levi that compares countries, their populations and GDP.

Whether any particular method increases the collaborative intelligence of distributed teams can only be demonstrated in practice. For some data, such as World Population or World Population Growth Rates or Annual World Population Change merely seeing the data rendered visually is a stimulus for problem-solving. But the more interesting and challenging question lies in how to visualize the problem-solving process so that all agents participating can increase their collaborative intelligence. The TRACE Cognitive model proposes one representation of process steps.

FUTURE DIRECTIONS. Ecosystem forecasting and sustainable remediation is an ideal testbed to integrate Decision Support Systems that can later be replicated elsewhere. Because understanding each sustainable remediation problem requires visualizing its complex system of variables as a whole, with all of its physical components dynamically linked and naturally constrained, an integrated decision support framework needs capacity for

• Linking models to simulate alternative impacts given indicator changes. Once calibrated, the model, together with other linked models addressing many indicators in a gven problem, can be used for Ecosystem Forecasting and for projecting “what if?” scenarios for given sets of variables and growth scenarios. Future capacity for scenario-building will allow planners to try out alternative hypotheses and associated decision sets to observe how scenarios play out in a simulated environment and explore outcomes, intentional or not. 
  
  
• Animation of alternative urbanization scenarios. Data can be used to construct an animation that shows the extent and pace of urbanization in the region. Computer models of urbanization calibrated with historic data will be useful for Ecosystem Forecasting, projecting changes for given sets of variables and growth scenarios.
  
  The Data City Network initiative is committed to fostering greater awareness of the impacts of urbanization.
  
  
• Multi-scale views: in and out zooming. Integrated planning requires capacity to see systems at all relevant scales. The challenge is to recognize ranges-of-scale relevant for each planning/ management question through integration-modeling methods and advanced GIS spatial analysis and visualization techniques.
  
  
• Knowledge domain visualization. Beyond the organization of geo-spatial data and its links, knowledge domain visualizations give users a way to identify research areas, experts, institutions, grants, publications, journals, etc. in their area of interest. In addition, they can assist to identify interconnections, the import and export of research between fields, the dynamics (speed of growth, diversification) of scientific fields, scientific and social networks, and the impact of strategic and applied research funding programs etc.
  
  
• Process maps: status of collaborative problem-solving. Visualization of problem information is seldom adequately complemented by visualization of the problem-solving process. Information visualization goes beyond viewing static data to enable users to collaborate and to control and update the environment. It offers simple ways to navigate through information so the user can see where he and other users are on a map of their collaborative problem-solving process. This diagram illustrates one example, developed by Zann Gill to support a Smart Cities initiative.
  

to top | home | links | compass