If Congress wants Federal agencies to get better at big data collaboration, they should start by fixing the outdated Privacy Act, according to Jessica Kahn, director of the data and systems group at the Center for Medicaid and Children’s Health Insurance Program Services.
The Privacy Act of 1974 requires Federal agencies to give the public notice of their systems of records by publication in the Federal Register. However, Kahn said that the legislators who wrote the act could not imagine the technology that would be prevalent in 2017, and that policy should shift to reflect the current technological landscape.
“The fact that we’re talking about a law from ’74 is inherently challenging. A lot of what raises people’s concerns about privacy can be addressed with tech,” Kahn said. “It’s no small feat for us to arrange data transfer. We should have data exchange first policy.”
Kahn, who spoke at Cloudera’s Government Forum on April 25, said the Privacy Act should be updated to include data exchange-centered policies. Kahn’s office collects data from 56 states and U.S. territories and tries to “make them work and smell like one data set.”
One of these partnerships is the Federal Communications Commission’s Lifeline program, which offers discounted cellphone use to people who are enrolled in Medicaid. The FCC relies on Medicaid data to verify who is enrolled for health care. This program marks the first time Medicaid is working with the FCC to create a data agreement, according to Kahn.
“We’re only now getting a handle on open data sets,” Kahn said. “Harmonizing data has been a challenge.”
Kahn stressed that the Federal space offers an important mission; she compared these government positions to an “internal Peace Corps.” Jobs at startup companies are attractive to recent college graduates because they often offer more money than Federal agencies.
She also said agencies could benefit from establishing strategic data management plans early on and designing infrastructure with users in mind.
“It’s often an afterthought. It should be front and center,” Kahn said. “What drives me bonkers is when someone comes to me with an IT solution and then asks me if I have a problem. Start with data users, build use cases, and then build infrastructure around it. Please don’t email me and shop me with data projects. I feel pretty evangelical about this because I work in a large government agency with many people working on data and they don’t always talk.”
Interagency communication was another major theme addressed at the Cloudera forum. Ensuring that all teams within an organization are on the same page is key to gleaning insights from big data, according to Peter Kovac, manager of high performance computing and data engineering at the Department of the Treasury’s Office of Financial Research.
Kovac said projects fail when too much emphasis is on the development team and the security team is left behind; he recommended involving business security people early on in projects.
Jeffrey Thomas, information security program manager at the Johns Hopkins University Applied Physics Laboratory, said that his teams represent different aspects of the university’s IT department. He compiles his project groups by gathering people from different areas of the IT organization to create one big team.
“Communication is critical. It’s new to folks even in IT space,” Thomas said. “You always have to be out there and keeping people on the ball.”