Small Data Can Accomplish Big Changes

There’s plenty of talk about big data and how it is changing everything, including education. However, colleges and university officials should be thinking about “small data,” according to Stephen Laster, former chief information officer of the Harvard Business School and current chief digital officer of McGraw-Hill Education.

“I love the idea of bringing data together to look for patterns and insights, but not in education,” he said during his keynote address at the Campus Technology 2014 conference in Boston.

A lot of students start school, but not enough finish, a fact that Laster considers “almost a crime.” A program that helped summer students at the University of Texas El Paso prepare for college math through the use of knowledge maps and artificial intelligence engines is just one example of the kind of small-data initiative Laster recommends.

“Small data is not going to be the sole driver, but it can be one small piece,” he said. “We cannot unilaterally drive the change, but we can drive a lot of it. Let’s start small.”

Higher Ed Should Beware of Hackers

Colleges and universities across the country may be making it much easier for hackers to steal sensitive information from students and their parents, according to research done by Halock Security Labs.

A survey of 162 institutions found that more than half of colleges and universities transmit sensitive information over unencrypted channels, including financial statements. In addition, one fourth of the schools said they ask that personal information be sent by e-mail.

“When universities utilize unencrypted e-mails as a method for submitting W2s and other sensitive documents, the information and attachments are transmitted as clear text over the Internet,” Terry Kurzynski, partner at Halock Security Labs, told eCampus News. “This format is susceptible to hackers and criminals who can use this private information for identity theft.”

The report said the open culture of higher education and budget issues facing colleges leave IT departments without the funds to protect student information. It said campus administrations may not completely understand the dangers of sending the information over unencrypted channels, but should since the issue could draw the attention of federal and state government agencies.

“These are foreseeable risks that are extremely treatable,” Kurzynski said. “Breaches resulting from this type of transmission will capture the attention of states’ attorneys general and the Federal Trade Commission.”

Big Data's Power Lies in Profiling

“Big data” is one of the hot buzzwords circling the business, technology, and education sectors these days. It’s not really a new concept—the smart players have been collecting and crunching numbers and data points all along—but system enhancements are making it possible to gather more data than before and more people are becoming aware of the power of data analysis.

In his Thought Leader educational session at the Campus Market Expo (CAMEX) 2013 in Kansas City, MO, Michael Wesch offered an anecdote that illustrates that power. He is the associate professor of cultural anthropology, Coffman Chair for Distinguished Teaching Scholars, Kansas State University, Manhattan, and U.S. Professor of the Year.

Wesch told the audience that a colleague became incensed when Target sent his 15-year-old daughter a mailer touting pregnancy and baby products. The colleague complained at the local store. A few days later, he returned to apologize. “It turns out Target knew the daughter was pregnant before he did,” Wesch explained.

Target didn’t actually know that, but had surmised the situation when the daughter bought two particular products at the same time. Purchased separately, the products meant nothing. But purchased together, according to Target’s data profiling, indicated there was a high probability the consumer was expecting. The dual purchase automatically triggered Target’s targeted mailer.

The anecdote shows that big data involves more than simply counting up consumer actions. The real value is in correlating data to create a profile of future actions.

In the education world, Wesch noted, software developers are trying to build similar profiles of student learning behaviors. He cited a couple of examples: Learning management systems, such as Blackboard, are exploring the huge potential in the big data they collect on users. The Knewton adaptive learning platform employs data analysis to guide learners through progressive levels of questions.