MARANA, Ariz. -- The truth is, we don't know much about big data.
We know what it is; sure. (Well, if you cut through the marketing hype.) And some of the more technically inclined among us know how it works. But what about the possibilities? Right now, it's too early to be absolutely sure. We can guess, but we just don't really know yet. It's too early.
Industry leaders gathered at the Techonomy conference here in Tucson to daydream a bit and -- well, true to the panel discussion's idiomatic title, see the forest for the trees.
How can digital pattern detection help the worlds of finance, retail, government, energy, healthcare or science? QlikTech' Lars Björk, Factual's Gil Elbaz, TIBCO's Vivek Ranadivé and photographer and The Human Face of Big Data author Rick Smolan gathered to answer just that. (The Harvard Business Review Group's Justin Fox moderated.)
Their conversation careened from topic to topic, but there were a few nuggets of insight to be had.
"What if you put all the emphasis on the user of the software?" Bjork asked. "What if you make a piece of software that works the way your brain works?" Not hierarchically, but associatively. How would that change things?
"A little bit of the right information beforehand," Ranadivé warned, "is much more valuable than all the information six months after."
Big data can help find patterns where we as humans may not see them. "If you buy Champagne, razor blades, and diapers -- it's probably a stolen [credit] card," he said, because diapers can actually be resold. To us, they pass the purchase off as legitimate, because the buyer looks like a parent.
"The more data that you have," Elbaz said, "the better the model that you will be able to build." He added: "Sure, the algorithm is important, but whoever has the most data will win."
Take the global economic downtown, for example. "The world is running in real time," Ranadivé said -- economic forecasting needs to, too. Consider: if you ran your home's HVAC system the same way the U.S. Federal Reserve runs itself -- in terms of big data intelligence -- "you'd constantly be overheating and underheating your home," Ranadivé said. Because the Fed doesn't know things in real time; they know it three months later.
Closed-loop mechanisms and making adjustments along the way are not new. "That's how you fly a Boeing 747," Ranadivé said.
Healthcare stands to be disrupted by big data very soon. "We're fast approaching that point where we're going to have extremely personalized medicine," Ranadivé said.
But there are new issues to confront with that. Say you're a patient in hospital, and you request six weeks of heart data that has been collected during your stay. The company refuses, saying "It's our data." Well hold on, wait a minute -- it's the patient's heart to begin with, Smolan said. So who owns the data?
Big data can tackle some of humanity's largest problems: pandemics, power outages, airplane malfunctions all can be prevented. "With data, we have the potential to eliminate food shortages," Ranadivé said. "We've spent a lot of money on technology. Now you know what [basketball player] Shaq had for lunch, but you still have food shortages."
Nonetheless, we can't lose sight of what is -- and isn't -- big data. "What is the definition of 'big data'?" Elbaz asked. "'Big' -- not just in volume, but variability [in terms of structure] and velocity."
Bjork agreed. "Big data: is it a relative term or an absolute term?" he asked rhetorically. "Big data is big to most people.
"The big data discussion needs to be put into perspective."
Photo: Asa Mathat