IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.
Sponsor Content
What does this mean?

Cloud Enables Platform to Help State and Local Governments Become More Data-Driven

COVID-19 highlighted the importance of using data more effectively, and the platform states and localities can use to do so already exists.

Shutterstock_Cloud Tech_Hand
Like many states, the Illinois Department of Employment Security processed more unemployment insurance (UI) claims in the first three weeks of the COVID-19 pandemic than all of 2019. But because the department is a member of the Administrative Data Research Facility, which runs in AWS GovCloud (US), leadership and agency staff were able to work with partner universities to quickly visualize the geographic dispersion of UI claimants, analyze their demographic characteristics, determine the economic impact on each county within the state, and view changes over time. By making data work for them, the department responded to citizen needs quickly and effectively.

“COVID-19 really forced the hand of many [state workforce agencies] to demonstrate agility and the production of timely and substantively relevant data, not only in Illinois but across the country,” says George Putnam, labor market information director at the Illinois Department of Employment Security. “That was critical because this crisis was unprecedented in its fluidity.”

There is more data than ever available to state and local government agencies today. But the public sector has thus far trailed the private sector when it comes to using data effectively. It doesn’t have to be that way, says Julia Lane, co-founder of the Coleridge Initiative, a not-for-profit organization working to accelerate data-driven work in the public sector, which spun off from New York University (NYU) in 2020 (Lane is also a professor at NYU’s Wagner Graduate School of Public Service).

“The private sector’s data revolution, which creates new types of data and new measurements to build machine learning and artificial intelligence algorithms, can be mirrored by a public sector data revolution characterized by attention to counting all who should be counted, measuring what should be measured, and protecting privacy and confidentiality,” says Lane.

Building a platform for better data sharing

In March 2016, Congress passed the Evidence-Based Policymaking Commission Act, which formed the U.S. Commission on Evidence-Based Policymaking, a 15-member agency charged with examining how government could better use its existing data to provide evidence for future government decisions.

“The overall goal was to make government become more efficient in its use of data to deliver programmatic services to citizens at lower cost while protecting privacy,” Lane says.

Achieving that goal meant ensuring data could be accessed and used in a highly secure, scalable, and cost-effective technology platform. The Office of Management and Budget charged the U.S. Census Bureau with building such a platform to inform the decision-making of the Commission. At the request of the Census Bureau, a team of data scientists at NYU, the University of Chicago, and the University of Maryland worked together to build that platform and called it the Administrative Data Research Facility (ADRF).

“We determined the best way to build the platform we needed was to use the cloud. Cloud would enable scalability and security while keeping our costs down,” says Lane, who is also author of Democratizing Our Data*, a book that explores how to build data systems at the state and local level.

The team chose Amazon Web Services (AWS) to build the platform, which made sure it could provide state and local government users a FedRAMP-certified architecture that would enable them to host confidential data and facilitate data sharing.

The cloud had an additional benefit for state government agencies because it enabled them to understand how their state fit into the activities of the region. State data systems, unlike state activity, end at state borders, so as a result, each governor or state agency is operating without sufficient information to make informed decisions about, for example, the returns to state investments in education.

“Workers easily move across state lines to get work, but states have not been able to easily share knowledge about those movements,” Lane says. “But once states realized they could securely host their data in the cloud, control it, and monitor how it is being used, it was a game changer.”

The eventual result was a fast and secure computing platform designed to support federal, state, and local government agencies that wish to share and analyze datasets from confidential sources to support state and regional decision-making.

Putting the platform in action

The next challenge was building capacity in the use of administrative data — data generated using government administrative programs. The NYU, University of Maryland, and University of Chicago team formed a partnership to develop and deliver Applied Data Analytics training programs to teach participants analytics skills and how to apply them to their own real-world data to solve state-defined problems. That training program has since been expanded to enable state universities across the country to deliver the training to agencies in their states using ADRF as the training platform. The university/state partnership — which was inspired by the success of the Morrill Act, which provided grants of land to states to finance the establishment of colleges that focus on education, research and extension services — enabled over 600 government agency staff from over 100 agencies across the country to be trained and to generate new approaches to solving state problems.

Recently, several Midwestern states, including Illinois, formed the MidWest Collaborative, which is designed to use both the training programs and the ADRF platform more fully.

“The participation of the states and their ability to drive a collaborative agenda has resulted in a data-analytic synergy to support evidence-based practice around shared policy imperatives,” says Lane.

And those ideas bore fruit. Although the ADRF platform was completed prior to the COVID-19 pandemic, it proved invaluable as the pandemic took hold.

As luck would have it, the MidWest Collaborative met in Columbus, Ohio, in March 2020 to identify potential joint products and data models. When the pandemic hit, “many state employees across the U.S. were sent home in the early stages of the pandemic and had no access to their data. But states that were members of the Coleridge Initiative at that time could still access their data in the cloud,” says Lane. “Then they realized they could use it to understand the situation that was unfolding better, like Illinois did. We ingested and analyzed Illinois’s UI claims data using ADRF. State leaders were then able to assess the situation very quickly and with a relatively low burden on the state.”

Data is the new gold

The Coleridge Initiative team is currently investigating technologies like artificial intelligence and natural language processing to see how they might be used to further improve the organization’s work and members’ use of data.

Going forward, Lane says the organization plans to continue to work with state agencies and universities to determine how to structure its training programs and foster state and local government data use and collaboration.

“COVID-19 laid bare the inadequacies of our current data system,” Lane says. “It was a wakeup call. We can and must do a lot better. The state agencies and universities have shown that the ADRF can be used to do so.”


* Democratizing Our Data: A Manifesto, https://mitpress.mit.edu/books/democratizing-our-data