Better data management makes digital services shine
Leveraging metadata and establishing data sharing frameworks across agencies can help government achieve higher data quality and security.
Government should adopt data practices that put organization, security and quality at the forefront, state officials said.
A strong data management foundation can help government more efficiently share information across departments and provide the public with data that promotes self-sufficiency, Dessa Gypalo, chief data officer at the Illinois Department of Innovation and Technology, said at a Jan. 31 webinar.
Gypalo was joined by Oregon CDO Kathryn Helms and Arkansas’ Department of Information Systems Chief Analytics Officer Robert McGough at the online event hosted by Data Labs, a venture between the Beeck Center for Social Impact and Innovation and the National Governors Association.
The experts cited metadata as a valuable tool for boosting agencies’ data quality. For instance, during Gypalo’s time as CDO for Cook County, metadata standards were applied to open data released to the public, she said. Two major benefits arose from this practice. One was a reduction in Freedom of Information Act requests as residents could answer their own questions because “the ‘data about the data’ gave people the information they needed,” she said.
Metadata also gives departments a roadmap for increasing data quality and integrity. It gives data users a better understanding of where the information is coming from or when it was last accessed, which may reveal where data collection needs improvement, Gypalo said.
To foster effective data organization, agencies should have a common goal that justifies the need to share data points across departments. For instance, the data sharing agreement for the Illinois longitudinal data system—which tracks the outcomes of students as they progress from pre-K through postsecondary education and into the workforce—covers eight state agencies. With the goal of reducing unnecessary data exposure, officials from those agencies regularly discuss why certain data points should be shared and what data is needed to answer their particular questions, Gypalo said.
“I'm a case manager, but I find that I don't necessarily always have access to the data or information about the specific client that I'm working with. Or I might be working in a workforce development panel, but I don't have enough intel about the other services that my users are receiving,” Gypalo said. “It's not just your databases that need to be relational. It's also the conversations you have with your end users.”
Data dictionaries are another tool that can help agencies polish their data management. For example, Oregon agencies are encouraged to outline any dataset limitations such as identifying margins of error in budget information, Helms said. Another way to bolster data organization is intuitively naming attributes so that data fields remain “clear and unambiguous” when building or interacting with data systems, McGough added.
Data inventories and data catalogs are crucial factors in maintaining security and privacy, Gypalo said. By establishing these systems, it’s easier for officials to identify what legal protections apply to their data and then share information appropriately.
Helms also recommended a tiered governance structure to ensure data sharing does not compromise individuals’ security and privacy. For example, Oregon’s state longitudinal data system is led by a governance committee that identifies and prioritizes research initiatives and then an executive committee that provides feedback on proposed research questions to ensure data is collected and used properly. The state is also exploring the opportunity to create a chief privacy officer role to help deploy enterprise privacy approaches, she said.
“States and government in general tend to be more risk averse than they need to be … with some of these data quality issues,” Helms said, but implementing accountability measures such as metadata standards and data sharing and management frameworks may alleviate agency concerns.
Data Labs—a no cost training and technical assistance program that helps states launch data sharing projects to improve government services—is taking applications for its next cohort. Proposals are due Feb. 15 and an information session will be held Feb. 3.