Mining Big Data
Did you hear the one about the teenager in Minneapolis whose parents freaked out about the baby-related coupons Target was sending her?
Turns out, she really was pregnant; she just hadn’t gotten around to telling them yet. Target sent the coupons based on the fact that she was buying things such as unscented lotions and vitamins, things the company’s data indicated pregnant women tend to buy.
It sounds like science fiction, but it’s the reality of so-called “big data” and the level to which it allows companies to market to specific customers. In recent years, the size of datasets available have multiplied exponentially, and they’re still very cheap to store.
In conversations with professors in the University of Iowa’s Tippie College of Business, executives from companies such as Target, John Deere and State Farm Insurance said they want students trained to organize and analyze their companies’ data and use it make predictions and suggest strategies. In response, the college has created a new undergraduate degree program intended to do just that. Enter, business analytics and information systems.
Businesses long have used data to make decisions. The difference now is the explosion of data that’s available — one theory suggests the amount doubles every couple of years — and the ability of even small companies to process large amounts of data at a relatively low cost, said Gautam Pant, a Tippie associate professor of management sciences.
“It requires a small computer, not much hard-drive space — so it doesn’t cost much,” he said. “Now the question a small business may think about is, ‘What kind of questions should we ask from the data?’”
From there, it takes a more creative mind to devise a mathematical or quantitative approach to use that data to make predictions, said Jeffrey Ohlmann, Tippie associate professor of management sciences. While there certainly is a scientific method to working through the numbers, it takes a human to apply that to problems they see in the real world, he said.
“Often people view this as you dump in the software, you turn a crank, turn it a few times, out comes the answer and then you just have to be able to understand it,” Ohlmann said. “Really, there’s kind of an art form to it.”
Data also can be humbling because it can trump intuition, Pant said. It can draw from millions of observations and prove a person wrong millions of times, he said.
“Big data can prove you wrong big time,” he said. “That’s a humbling experience. It’s true failure you can learn from. That’s a very good learning experience.”
Humans tend to favor simple answers to questions, which often drives them to fall upon conventional wisdom, Ohlmann said. He called upon the example of an entrepreneur who struck it rich drilling oil wells using a different set of metrics to select drill sites than a massive oil company, which had been rejecting potential sites based on decades-old theory.
Companies will now be expected to use solid data to back up their positions, said Michael Hasler, who directs a new M.S. program in business analytics at the University of Texas at Austin.
“I think business and decision-making in general is trending away from going with your gut and more toward looking at the data,” he said.
UT’s one-year program, based in the business school, starts this fall. Hasler said it’s received significant programming support from enthusiastic companies such as Walmart, Dell and IBM. University officials projected 150 applicants and about 400 applied for the program, he said. Of those, 53 were admitted.
UI’s BAIS program, which also starts this fall, is one of the first undergraduate programs of its kind in the country. Experts say most big data-centered programs in higher education are offered at the master’s level.
A master’s-level program made the most sense for UT because Hasler said he wanted to attract students from a variety of undergraduate backgrounds such as physics, engineering and math in addition to business, which comprises about 20 percent of the students.
Ohlmann said companies have expressed interest in both students with undergraduate and graduate-level training. The BAIS program is designed to provide students with essential data analysis skills. A company could then hire them and train them on the nuances of their specific software, he said.
Other companies will look for graduates with specific master’s-level training, Ohlmann said. The Tippie College of Business is currently considering a master’s level program in big data, but it’s yet to be seen whether that will materialize, he said.