Think Big

Big data is the collective and constant answer to one question: What is happening in the world right now? There is no why or how, no cause or effect, just the simple observation of the present.

Big data encompasses all, but how is that information collected? Michael Schroeck, partner/vice president, IBM Business Analytics Optimization, and Viktor Mayer-Schönberger, co-author of the current business best-seller Big Data: A Revolution That Will Transform How We Live, Work, and Think, visited the Adelphi Garden City campus for separate lectures regarding the topic. Schroeck explained that his sector has never before experienced a coming together of so many information sources, including the cloud, mobile devices and social media.

Any and every kind of data can be and is collected. For example, Mayer-Schönberger said:

• People buy lots of Pop-Tarts at Wal-Mart before a hurricane makes landfall.
• Google processes a couple of dozen petabytes, each equal to a million gigabytes, every day; 10 million photos are posted to Facebook every hour; an hour of high-definition video is uploaded every second to YouTube, which has 800 million users; and 500 million tweets are posted on Twitter every day.
• Used cars that have the fewest defects are orange.

“Information is everywhere. How can we, as an industry, do a better job not just tapping the information, but refining it, analyzing it and turning it into something that’s very useful?” Schroeck asked at his recent campus lecture. It’s the job of his department to tap into the world’s unrefined information and to figure that out. “The world’s information doubles every 11 hours. Ninety percent of the world’s total information from the outset of history has been generated in the last two years, and it’s not going to slow down.”

Mayer-Schönberger used the Sloan Digital Sky Survey telescope as an example. In 15 years, it has collected more than 200 terabytes of astronomical data. He explained that the 8.4-meter Large Synoptic Survey Telescope is slated to start sky scanning in 2016 and will collect those same 200 terabytes in about five days.

He cited two other examples:

• When, in 2003, the scientific world celebrated the first entire mapping of a single DNA sequence, he noted the achievement took 10 years and a billion dollars. Today, 13 years later, it can be done in two to three days and at a cost of less than a thousand dollars.
• It’s due to big data—the availability of massive amounts of data—that scientists have discovered patterns in the vital signs of premature infants to predict, with a high degree of likelihood, the arrival of complications 24 hours before symptoms manifest.

“We’re datafying ever more aspects of human reality and then we extract value from it,” Mayer-Schönberger said, noting Google Glass as one such item that could turn human reality into a constant stream of information. “To be able to datafy the human gaze—that is, to know, to find out and to be able to observe what human beings are actually looking at—can you imagine how valuable that is in terms of advertising dollars when we really find out what kind of adverts work because people are actually looking at [them]?” he said. But it’s these kinds of prophecies that tend to scare the populace, and for good reason.

In 30 U.S. states, Mayer-Schönberger said, the decision on whether a prisoner is paroled is being made in part by doing a big data analysis on the likelihood of that person’s committing another crime within 12 months. Predictive policing is also in use.

Mayer-Schönberger explained that using big data in such ways is not the best application of the data, because the collected data can’t explain the whys. The data can only observe. “The problem here is not big data. The problem is that a correlation is not causality, that why and what are different, that what doesn’t tell us why. If we try to tease out [causes from] correlations…and therefore hold people responsible, we’re falling prey to something of an abuse of data.”

So what’s next? Many people think of the Hollywood blockbuster Minority Report, but Mayer-Schönberger pointed out that if humans remain the masters of big data, then the future will reveal humans who are smarter and who can make better decisions.

“As much as it is vital to learn from the data, we must also preserve a space for the human.”