Aneesh Chopra became the nation's first chief technology officer in 2009. Much of his work focused on the use of big data to inform big decisions. He noticed something then that he continues to see today: When it comes to making major decisions, there are two camps. One consists of people who believe intuition trumps analysis—go with your gut. The other rejects intuition in favor of careful data analysis—where there is enough data, there's no need for intuition.
Chopra, A&S '94, has come to see this as a false and unproductive divide. "The ideal is a marriage of the two," he says. "You want to foster more creativity and intuitive thought, but you also want to test hypotheses rapidly with the benefit of the horsepower that comes from the big data movement." The latter requires knowledge of the massive data sets becoming ever more available, as well as lower-cost cloud-based analytical tools to make wise use of them.
Chopra left the federal government in 2012 in an unsuccessful bid to become Virginia's lieutenant governor. After that, he saw a private-sector opportunity based on the need to understand smart use of data in decision making. So last October he founded Hunch Analytics, an incubator—Chopra seems to like the term "hatchery"—intent on helping health care and education providers mine data to make better decisions. Its chief executive officer is fellow Johns Hopkins alumnus Sanju Bansal, Engr '90 (MS).
Large corporations have the resources to invest heavily in developing data analytics. Smaller institutions do not. "It might be one thing for Wal-Mart or Sears or Facebook to spend ungodly sums of money on their data analytics infrastructure," says Chopra. "It's quite another for your local schools or hospital." He envisions Hunch Analytics launching a portfolio of products and services to empower such institutions with sophisticated, lower-cost analytic tools.
In health care and education, he points out, outcomes data, often collected by government and made increasingly available, are becoming ever more important. For example, the Affordable Care Act has mandated a change from hospitals working on a fee-for-service basis to more of a fee-for-outcomes system. There will be a premium on making people well and keeping them well because hospitals will be reimbursed at a much lower rate for patients who bounce back within weeks of their initial treatment. A hospital like Johns Hopkins has internal data on everyone who has been admitted. But there's a vast trove of external data, Chopra says, that if properly analyzed can help management identify ways to keep populations healthier. For example, Medicare patients can download three years of their personal health care data through the Medicare website. This data set is called a "blue button file" because of the website button that beneficiaries click to obtain it. Chopra asks, "How many Medicare patients today have visited Johns Hopkins, and how many of them have proactively shared their blue button file? If Johns Hopkins had access to that file, how many more questions might it be able to answer? With permission from its patient population to access such data from outside its system, it could perform analytics that might indicate who, out of a million people in the Baltimore-Washington region, are the people they should be most worried about today, whom Johns Hopkins should be reaching out to and encouraging to live a healthier life."
Posted in Science+Technology
Tagged big data, data mining, analytics, hunch analytics