When an engineering organization like the IEEE gathers some of the leading lights of big data for a day-long symposium at Silicon Valley’s Computer History Museum, you might expect the topics to be deeply technical. To be sure, Rock Stars of Big Data featured plenty of that, but the speakers also focused on the cultural, personal, and even ethical issues surrounding Big Data.
They also found time to marvel at the fact that data scientists had somehow become “sexy!”
Here are the Big messages I took away from the Big Data day.
1. Big Data is Not a Technical Issue
“For 95 percent of all problems, the technology and tools are available to handle everything you need to do,” claimed Mark Seager, Intel’s CTO for Technical Computing. Instead of technology, it really comes down to people, politics, and policies. “The technology is there to implement,” he said, “but you need buy in, you need the talent, you need the budget.”
“Big data culture today is really in the stone age,” added Josh Greenbaum, principal at Enterprise Applications Consulting. “The technology has gotten way ahead of what we can do culturally. We’re good at cranking out MBA spreadsheet jockeys who don’t get data. They understand how to manipulate data, but they don’t understand it fundamentally.”
2. Big Data is About People
“When there’s a problem with a Big Data implementation, it’s usually a not a systems problem but a people problem,” said William Ruh, vice president of GE Software. “People don’t want to have the way they work scrutinized,” he said. In failed Big Data projects, the people affected often were misread or ignored.
For Ruh, the key to outcomes that matter is tying Big Data into people’s real agendas. “If you leave the person out, you will not solve the right problem,” Ruh said. “The best analytics are ones that gives you insight into who to talk to.”
3. The Chicken or the Egg? The Data or the Question?
One view of Big Data holds that organizations should gather as much information as possible. “When you start to dig into the data, you see things,” explained Michael S. Johnson, director of utility for care data analysis at Kaiser Permanente. “You can find things even when you don’t know you’re looking for them.”
But Bill Franks, chief analytics officer for Teradata, thought that view was backwards. “You should not be collecting masses of data in hopes that it will solve a business problem,” he said.
“Big Data by itself is kind of useless,” added Chris Pouliot, director of algorithms and analytics at Netflix. “You need big analytics to make a big impact.” Besides, as Greenbaum pointed out, you don’t really even need Big Data for big analytics. “Lots of important findings come from relatively small data sets.”
4. Big Data Can Change a Company’s Business
Nike is known for fashion-driven shoes and sportswear, said Teradata’s Franks. But its FuelBand fitness monitor is a high-tech device, connected to a downloadable app that stores, tracks, and analyzes the wearer’s activity data. That puts Nike in the high-tech manufacturing business, the data storage business, the Web and app software business, and the Analytics-as-a-Service business.
The success of the Fuelband is based not on style, Franks said – fitness trackers are all pretty ugly – but on the quality of the analysis it delivers to the wearer.
5. Big Data Lasts Forever
Grady Booch, chief scientist at IBM Research and co-founder of Computing: The Human Experience, called Big Data “a genie without a bottle,” because “information once made digital can never be called back.” No matter what legal or other protections it may appear to have, he said, at some point someone you may not trust could have access to that data, because data’s lifecycle extends beyond our control.
6. Collecting Big Data Can Create Big Liabilities
As our technology outstrips our laws, said Booch, key questions have yet to be addressed: Who has the real ownership of a given data set, technologically, legally, and societally? Who takes responsibility when agreements are violated?
We simply don’t have the body of law to deal with those questions, and they are not moot. Like many other Big Data collections, the U.S. Census is protected by law from being used for any other purpose. "... but laws can be fungible,” Booch noted. Those laws didn’t stop the government from using Census data to identify Japanese Americans before sending them internment camps during World War II, for example.
On a practical basis, improving your data collection and analytics may mean being held to a higher standard of quality, warned Greenbaum. “Just because you collect the data, does that mean you are responsible for analyzing it?” asked Teradata’s Franks. “This is uncharted territory,” warned Ruh. “What are you required to keep and what are you required not to keep?”
7. With Big Data Comes Big Responsibility
Beyond the practical considerations, technology professionals have a responsibility, Booch said, to be cognizant of the possible effects of the data we collect and analyze. “The big ethical issue,” he said, “is that nobody thinks this is an ethical issue.”
The consequences are very real. “I see it coming,” Booch said. “We will see some really sad, heart-wrenching uses of data that will destroy an individual, and possibly groups of people.”