Opinion: Spatial visualization is starting to catch up with fiction.
Considering how well our brains handle spatial relationships, its unfortunate that most information systems make little use of that ability to let us find patterns or anomalies in data.
I first wrote that sentence as "its surprising and unfortunate," but on reflection I had to admit that its not surprisingbecause, on the programming side, spatial representation and presentation are terribly difficult.
Back in 1983, my first technical project on joining a new employerwhich was just bringing in those still-newfangled IBM PCswas to craft a three-dimensional visualization of schedule, cost and technical risk. This was meant to provide an abstract view of a complex project that would let a project team get an immediate idea of where its critical focus needed to be.
Using a ridiculous amount of trigonometry, and conventions of perspective rendering that I had learned in 8th-grade drafting, I came up with something that looked like a three-dimensional abacus. It presented a 3-D grid of blobs whose sizeas I recallrepresented technical uncertainty and whose x, y and z coordinates represented cost, time and
something else. (Sorry, its been a while.)
This tool might have had potential, but a 4.77MHz IBM PC that lacked a numeric coprocessorat a time when few programming tools even supported such hardware without resorting to assembly languagewas simply not fast enough.
It took much too long to paint a new view of anything but the simplest demonstration data set. It was not possible for the user to look at that "risk cloud," twirl it around different axes, recast it in different perspectives and have the overall degree of interaction thats needed to make an on-screen object seem real to the brains perceptual machinery. Thats a crucial threshold to cross.
Before the end of that decade, though, I could take a mouse and move a hand-shaped icon to spin a 3-D data cloud in any desired directionand that was on a Macintosh. Plus, with an 8MHz CPU and 1MB of RAM, affording a fraction of a percent of the processing power and memory that were commonplace even before we celebrated Apples 30th anniversary on the first of April.
Clearly, the hardware resources are now at hand for far more powerful data visualizations, but theres still an enormous asymmetry between the brains ease of interpreting a display and the programmers effort in providing it.
I discussed the need for a smoother path from concept to delivery of spatial analysis and insight in a conversation in late March with Dean Stoecker, president and CEO of SRC. The company just released its Alteryx spatial database management tool, offering new insights from heterogeneous collections of spreadsheet or database records.
"We started in 1997 with a white-board and no product and no customers, so all we cared about was making it easy for anyone to process any data from any kind of data store anywhere on Earth," said Stoecker. "Our engines are platform-independent, database-independent, country-independent; some of our customers run in a browser, and all they care about is what comes out."
Thats exactly what I want to hear from a provider of analytic tools today, and its a set of criteria that was almost inconceivable when I was wrestling with trig formulas in 1983. Whats crucial, said Stoecker, is the realization that spatial techniques are only incidentally about geographical maps.
There are many useful questions that can be asked using maps, he observed, such as, "How many of my high-volume customers live within 5 miles of one of my retail locations?" Those questions are important, but spatial presentation can take many other forms: My favorite examples include those on display at cybergeography.org
William Gibsons coinage of "cyberspace" in the 1982 short story "Burning Chrome" set a high standard for visualization that the Net of today fails to meet, but tools like SRCs may raise our expectations. Maybe Ill even go back and finish the job I started in that other century.
Technology Editor Peter Coffee can be reached at email@example.com.
Check out eWEEK.coms for the latest news, reviews and analysis about productivity and business solutions.