by Fred Hapgood

VISUALIZATION – Seeing with Digital Eyes

Mar 15, 20026 mins
Data Center

Most days of the week we are as sober as judges, but we can be tempted. In April 1993 we proved the point by eloping with a looker by the name of visualization. “Soon…visualization will become the foundation for business planning, intuitive analysis, bilateral decision making and the creation of value,” we wrote. And not 10 years down the road either. “Vision’s time,” we announced, “has come.”

The promise that seduced us was the potential of developing striking, memorable, novel, visual metaphors for business processes and using those metaphors to make monitoring and managing those processes simpler, faster and more intelligent. If that could happen, managers would be able to push collaboration to new levels of complexity, work with much larger volumes of data and do intelligent sorting through huge solution spaces. We liked this prospect a lot. “Visualization tools are the most important information technology since the typewriter,” we said.

Nine years later it seems safe to say that this proposition still sounds a bit sweeping. True, the commodity end of data visualization?bars, pie charts, line graphs?did get a lot cheaper and easier, but pie charts don’t do much that can’t be done with text. We had something more ambitious in mind.

To be fair, those ambitions have come close to realization in a number of specialized sectors, including cutting-edge CAD programs such as Graphisoft’s ArchiCAD that allow building visualizations that architects can use for design, marketing for virtual reality tours, contracting for construction scheduling, and finally, when the building is complete, managing assets and facilities by tenants and agents. Farmers routinely build maps of their fields that represent chemical applications and crop yields on a yard-by-yard basis. Mechanical engineers use kinematic visualizations to test tolerances and clearances. Medical imaging is a continuing miracle.

However, it is worth noting that almost all this progress has taken place with data already possessing a spatial character. All visualizations are fundamentally metaphors: Each element has to stand for something. No matter how powerful a visualization might be, if that connection is not obvious to its users there is no point to the product. Pie charts worked because people came to them already understanding the difference between small and large slices of pie. Spatial data is inherently obvious in the same sense; it is easy to understand what it means to walk around inside a virtual building because we have so much experience walking around real buildings.

For visualization to be as important as the typewriter, the technology has to work where the spatial aspects of the data in question are either irrelevant or do not exist at all. These latter applications raise the old problem of finding a happy compromise between ease of use and power, between finding metaphors that are both easy to learn while delivering unprecedented levels of information density.

Some observers think the right path to visual computing is finding or designing new metaphors. Bob Jacobson, CEO of Modern Visualization based in Burlingame, Calif., suspects that inventing better visualizations might well depend on uncovering the fine details of how the brain does its own image processing and then generalizing from that vocabulary. He is hopeful that a new science called vernacular geography, the study of geospatial metaphors in everyday life (as in such phrases as “over the hill,” “around the bend” or “ain’t no mountain high enough”) will uncover the building blocks for this advance.

On the other hand, Tony Crescenzo, CEO of Illumitek, a visualization company in Herndon, Va., suspects another solution might lie in industrial-strength interactivity. He points out that if a system can transform complex data quickly enough, it doesn’t matter what the visual metaphor is (within reason), since the interactivity makes it easy to match data with onscreen objects and therefore makes that metaphor easy to learn. Higher interactivity also makes it easy to extend the power of familiar metaphors.

Visualization in the sense of an ad-vanced interface, however, makes sense only when it addresses very large amounts of data. (Simple, cheap and familiar representations?bar charts, for example?exist in abundance for smaller data sets.) For most of the past decade the costs of the storage of even a modest number of gigabytes and the computational resources required to manipulate them quickly imposed a price level that was prohibitive for all but the most value-laden applications.

In addition, until recently data compilations often entailed pulling scattered and usually incompatible databases together, adding programming costs on top of that. Crescenzo thinks that only now, when it is practical to think in terms of clusters of 2GHz processors running on top of data warehouses containing tens of terabytes, is the environment right for real progress in the art. (In addition, many companies now use storage networks that can pool their data with just a couple of mouse clicks.)

This emphasis on interactivity is leading some engineers to look for visual metaphors in the most highly interactive sector of the IT culture?computer games. “In practice, visualization is mostly about presenting changes in complexity that vary with time,” says Stephen Eick, CTO of Visual Insights, an e-business performance company based in Naperville, Ill. When his company started to design its products, the employees asked themselves what experiences their target market had with representations of that sort. The answer would have been obvious to any CIO, especially on Friday night, during the interdivisional Quake tournaments.

Eick suspects that the right vocabulary of commercial visualization will be very gamelike, rather like the vision of the matrix in William Gibson’s prophetic Neuromancer. Users, individually or in collaborations, will hover over data landscapes, zip down or up through levels of resolution, and transform them with a gesture. (Visual Insights has such faith in that vision that it built its products on the Microsoft Game API; in theory, the products could run on the Xbox.)

With some extrapolation, one could imagine that competitors will appear as the zombie reptiles from Xork, to be dispatched by lethally innovative product designs. Maintenance efficiencies will add or subtract vitality points. Players can spend procurement savings on magic armor or gas arrows. Promotion ladders will weave in and out of virtual environments?today a vice president, tomorrow, a wizard of the first rank.

This may sound like a trick to keep employees working 80 hours a week, but before we get upset let’s first see if anybody complains.