Higher Education’s Big (Data) Bang: Part OneAt a conference in October devoted to exploring the perks of Big Data in higher education, the event’s keynote speaker had a surprisingly contrarian take on the subject. Universities are investing millions in Big Data. “Big Data is bull—,” Harper Reed, the chief technology officer of President Obama’s 2012 campaign, said to an audience that included many campus IT officials hoping to learn more about Big Data’s benefits. Reed, who used the power of data to help Obama secure reelection, said the term is just a marketing tool meant to drive college and university IT officials toward expensive technologies for storing and analyzing data. If true, it’s a PR move that’s working. The University of Rochester has spent more than $100 million on “Big Data research.” Indiana University spent more than $30 million for a building to house its $7.5 million Big Data-crunching super computer called Big Red II. The Gordon and Betty Moore Foundation and the Sloan Foundation have pledged $37.8 million over five years to the University of California at Berkeley, the University of Washington, and New York University for a Big Data collaboration. If there’s any current set of buzzwords that rivals the popularity of massive open online courses (MOOCs) in higher education circles, it’s Big Data. But the phrase’s definition has a more elusive quality to it than the easily marketable free online courses. Just what is Big Data, and how are campuses using it, exactly? “Big Data is this exponential increase of information that’s been going on since the 1950s,” said Jim Spohrer, the director of Global University Relations Programs at IBM, a company that has partnered with campuses to drive the study and adoption of data analytics. Every decade since then, Spohrer said, has seen a certain section of business or society become a data leader. Maybe it was insurance companies in the 50s, he said, or the Sabre computer system for the travel industry in the 1960s. Spohrer refers to the information collected by these entities as “small data.” The small data has continued to pile up over the decades as information increasingly went digital, before finally exploding in growth in recent years due to online interactions. Social media posts, online news articles, digital scans of academic journals, student financial aid profiles, online medical histories of patients, transportation system sensors – taken altogether, this is Big Data. “We’re really at this stage in the last two years, where more data has been produced than ever before in human history,” Spohrer said. “It has become a new natural resource. An amazing natural resource.” Spohrer and IBM, of course, would be the spin doctors Reed referred to in his keynote speech. IBM now has more than 1,000 university partnershipsrelated to Big Data, including with Georgetown University, Northwestern University, and the University of Missouri. In January, IBM donated their supercomputer called Watson, which famously beat long-standing champs on the game show Jeopardy in 2011, to Rensselaer Polytechnic Institute. Watson has 15 terabytes of memory and can read 200 million pages of text in three seconds. With Big Data being comprised of so much information, it takes that kind of computing power to analyze and make sense of it all. Indiana University’s Big Red II, one of the 50 fastest University-owned supercomputers in the world, can crunch a quadrillion pieces of data every second. In order for a human being to perform the same level of calculations, it would require the person to complete a calculation every second of every day for 31 million years. The purpose of the Big Data that’s processed by these enormous and costly machines varies, but a primary one is helping universities continue their role as society’s go-to source of research and knowledge. “With the ability to process huge amounts of data at an almost unimaginable speed, they have become an essential tool in expanding the frontiers of knowledge, addressing the world’s most critical issues and probing the most fundamental questions about the universe in which we live,” Michael McRobbie, IU’s president and former CIO, said when the computer was unveiled. Universities also use Big Data technology to analyze their own student and faculty information — things like learning trends, financial aid statistics, and teaching costs. At Wichita State University, for example, this kind of data is used to predict which students will succeed at the college before they are ever admitted. Harper would call this “medium data,” but it is the kind of rapidly growing amount of information that learning analytics companies like Civitas Learning argue is also “Big Data.” With so many universities and companies buying into the idea of Big Data analytics, another use of the information and technology has come into play: training students to become “data scientists.” In two years, there will be an estimated 4.4 million jobs dealing with Big Data, but there are not that many students studying data science. Skeptics like Harper have dismissed this “Big Data skills gap” as a corporate-created non-issue, but others see it as a booming market for which universities have an obligation to prepare their students. Big Data degree programs since August have cropped up at George Washington University, University of New Haven, Bellarmine University, the University of Texas, and Northwestern University. “This sort of thing usually happens when there is some kind of market shape up and companies and universities don’t know how to even fill a position,” said Ellen Wagner, executive director of the WICHE Cooperative for Educational Technologies. “When that happens, people think ‘we’ll have to create a new degree for this.’” While universities that tap into this changing market early on do have an advantage, Wagner said, the job description of data scientists has yet to be fully defined. “There is this concern that there’s a lot of hype around data scientists without knowing what it even is,” she said. For some, that concern spreads to other facets of Big Data, as well. They worry that universities are sinking money into a murky swirl of hype and unnecessarily large amounts of information. A university recently crunched all of the “Big Data” it had gathered on a course and made a surprising discovery. Some in higher education question massive investments in Big Data. Out of the two professors who taught the course, one had significantly lower performing students. But this was a professor who had won several teaching awards and was well-respected by campus leaders. What was going on here, the researchers wondered as they sifted through all of the data points at their disposal. They could only draw one conclusion based on the data at hand: the poorer student performance was because the professor was not the level of teacher everyone believed him to be. The conclusion, it turns out, was wrong. All the calculations and algorithms failed to point out something most professors now know nearly instinctively: the course was at 8 a.m., and so the class was mostly made up of procrastinators who had waited until the last minute to sign up for the course. Hardly a room full of over-achievers. “When you have 5,000 data points, how do you know what question to ask?” said Sherry Woosley, former associate director of institutional effectiveness at Ball State University, as she related this real-life incident. That, she said, is one of the primary concerns as universities and colleges turn to Big Data technology to help make sense of all the information they have accumulated. More than 1,000 institutions are working with Big Data in some way, and universities are investing millions of dollars in super computers and data research centers. Critics like Harper Reed, the chief technology officer of President Obama’s reelection campaign, warn that universities are being romanced — by adaptive learning platforms or computer companies like IBM — into financially backing what will turn out to be a fad. Ellen Wagner, executive director of the WICHE Cooperative for Educational Technologies, said while she doesn’t consider the interest in Big Data to be a passing phase, the criticism does help to ground the conversation. “It’s great to remind us to put on our ‘reality glasses,’ to remind people that they need to be paying attention to the hype permeating this space right now,” Wagner said. “It’s really easy to talk about the ‘Amazonification’ of education. It’s easy to get seduced by what Amazon and these big transaction platforms do, but we have to remember that with education, we’re not talking about the same kinds of aggregated transactions.” It may turn out that the kind of data gathering that online retailers have used so well for product recommendations may not be so easily applicable to college course recommendations. Improving student outcomes is just one of many reasons a university may choose in invest in Big Data, however. Woosley said it’s important for a university to know what questions it’s hoping to answer with data before making the kinds of investments seen at Indiana University and the University of Rochester. A large research university hoping to uncover a common thread in thousands of car accident reports would be utilizing Big Data in a completely different way than a college just focused on helping students graduate. Too much data can be just as useless as not enough. Or even the same as no data at all. “People talk about information overload in terms of just email,” said Woosley, who is now the Director of Analytics and Research at EBI MAP-Works. “We’ll see the same thing with data reports. If you’ve given someone too much information, they’ll throw it all out as they don’t even know how to pick through it. There’s definitely a risk here of getting enamored with the phenomenon. Do you as an institution even need this huge data warehouse? The devil’s in the details.” Mark Milliron, Chief Learning Officer and co-founder of data analytics company Civitas Learning, said those details will decide an institution’s success with Big Data. Milliron said too many students are “flying blind” as they try to chart their higher education pathways, and Big Data tools are what will help fix that problem. “However, far too many data strategies end up as little more than expensive edu-voyeurism, focusing high-cost data work on accreditation, trustee updates, or little used reports that essentially watch if students succeed or fail,” Milliron said. “If that’s where the resources go, we doubt folks will see a return on the investment.” Jim Spohrer, the director of Global University Relation Programs at IBM, said Big Data is something people are just now beginning to understand, so the concerns and skepticism aren’t surprising. While he said he agrees that not all Big Data technology is for everyone – a community college may not need a supercomputer like IBM’s Watson – he cautioned those who dismiss the explosion in interest as just hype. “This is a real phenomenon,” he said. “Every job is being impacted by this. It really is time now to make sure all professions have some level of understanding of these tools. That’s what it’s really about, taking smarter actions, helping us make smarter decisions.” Woosley said Big Data’s role in higher education is not a black and white scenario. There are untold benefits to tapping into the wealth of information now available to universities, but there are some serious pitfalls as well. The knowledge of those pitfalls could get lost in the noise as the hype machine’s engine continues to rev. “I’ll admit I am starting to cringe whenever I see the term Big Data,” Woosely said. “It’s becoming this mythical thing, and that’s always scary.” The use of Big Data could unlock as much as $5 trillion in economic value a year, and it’s falling on colleges and universities to ensure such a boon happens. Big Data could unlock up to $5 trillion in economic value. Higher education institutions in recent years have invested hundreds of millions of dollars in data research and tools. Some universities are putting up this money themselves, while others are relying on government and organization grants. The Gordon and Betty Moore Foundation and the Sloan Foundation have pledged $37.8 million to the University of California at Berkeley, the University of Washington, and New York University for Big Data collaboration over the next five years. IBM has partnered with more than 1,000 institutions around the world to create Big Data seminars, courses, and even full masters programs.
The value for companies like IBM goes beyond just the obvious benefit of selling data tools to schools. The demand for people with data analytics skills is expected to increase by 24 percent over the next 8 years, according to the U.S. Bureau of Labor Statistics. In two years, there will be 4.4 million jobs dealing with Big Data. That’s far more jobs than people who can do them, said Jim Spohrer, director of Global University Relations Programs at IBM, calling the problem a “Big Data skills gap.” “From all sectors, there’s just a tremendous demand for it, so we’re working with our customers and stakeholders,” Spohrer said. “This is the kind of thing IBM does routinely, like in the 60s and 70s with mainframes. These kinds of new opportunities require new skills, and it’s a natural part of our business to do these private public partnerships to increase the number of people with those skills.” Ellen Wagner, executive director of the WICHE Cooperative for Educational Technologies, said she expects to see more and more institutions teaching data science to students, but it’s hard to tell if the workforce will be as large as IBM predicts it will need to be in just a few years. “I think you’re going to see places that believe they have a leg up on their fellow practitioners by offering these courses,” Wagner said. “It’s very hot stuff, but people aren’t sure what to do. Maybe it will be slow and start with simple courses, or it might be as full-blown as full degree programs.” If institutions can’t help close the skills gap, Big Data experts said that the global economy could lose out on trillions of dollars. A recent report by the McKinsey Global Institute estimated that “open data” alone (Big Data that is made freely available to others) would “unlock $3 trillion to $5 trillion in economic value annually across seven sectors.” The potential value of open data, the institute estimated, could be as much $1.4 trillion in consumer products; $920 billion in transportation; $580 billion in electricity; $510 billion in oil and gas; $450 billion in health care; and $280 billion in consumer finance. The sector with the most to gain, McKinsey Global concluded, is education. “Annual spending on K-12 and post-secondary education exceeds $4 trillion worldwide,” the report stated. “With so many resources dedicated to public education, there are substantial opportunities to increase the efficiency and effectiveness of current systems.” By using data to improve instruction, match students to programs and employment, make education financing more transparent, and increase the efficiency of system administrations, the education sector could see an increase of $900 billion to $1.2 trillion in additional annual value. All of this, of course, is reliant on schools actually making a push toward Big Data programs, Wagner said, and that’s why organizations, companies, and government entities are making a push of their own – that marketing move Harper Reed, the chief technology officer of President Obama’s 2012 campaign, called “bull—” in October. “I’m not opposed to what companies like IBM are trying to do,” Wagner said. “They have products that institutions want to buy, and they are all discovering that there is no winner if nobody is sure what to do with these amazing platforms and tools. The reality is that all of us should be more data-literate in post-secondary education, but we can’t just believe all the hype either. Everyone’s trying to figure out this very particular dance.” Follow Jake New on Twitter at @eCN_Jake, and join the conversation with #eCNBigData http://www./top-news/ibm-collaborates-with-1000-institutions-to-help-close-big-data-skills-gap/? |
|
|
