Ticker

6/recent/ticker-posts

Ad Code

Responsive Advertisement

Data-Driven Decision-Making Gets Real

Government agencies often espouse the need for data-driven decision-making, but where should the data come from? Who should determine how it’s used? And what kind of decisions are we talking about?

Taka Ariga, the Government Accountability Office’s (GAO) Chief Data Scientist and Director of its Innovation Lab, has many thoughts on these topics. He describes GAO as “the microcosm of the federal government, the candy store for data science,” and said the lab has a dual purpose: first, to understand how technology works so that GAO can analyze it, and second, to consider how GAO can use such technology for the agency’s own benefit.

Ariga believes the Federal Data Strategy is a good start — but his vision goes beyond that.

The Vision

“The vision I have for evidence-based policymaking,” he said, “is ready access to information, ready capability to transform that information into knowledge and ready ability to make decisions based on that awareness.”

He said a three-legged stool — consisting of data literacy, data governance and data science — makes a data-driven world possible. For example, a person must understand a dataset in order
to develop a narrative around it that makes the information more digestible to a broader audience.

“If people don’t consume and analyze what we put out there, why bother?” he said. “We need to meet the user more than halfway when it comes to developing an interactive product, and I think that comes after we ourselves have sufficient data literacy to interpret that information.”

To ensure that the data is worth looking at, an agency must follow strict data governance controls to confirm its quality. So despite a proliferation of public data on sites such as USAspending.gov, Ariga said GAO “categorically never relies on public data to perform our audits” because the agency can’t confirm its completeness or accuracy. And if you put great data on an unsecure server, that raises a Pandora’s box of other issues, he noted.

Regarding data science, Ariga said the problem you’re considering should dictate how you analyze the information. Perhaps statistical modeling is appropriate, or simulation exercises, interactive visualization or machine learning, among other options.

The Obstacles

Ariga said the greatest challenge is the general perception that information is power, and that sharing data could lead to a disastrous loss of control. He believes that sharing information is actually an opportunity for improvement, as long as agencies respect Health Insurance Portability and Accountability Act and other privacy constraints.

Regarding technology, the federal government’s computational abilities often are limited despite agency modernization efforts, and those limits can make it difficult to publicize information.

Data silos are a problem. For instance, Ariga pointed to the fraud surrounding pandemic- related social services benefits. “I think we all acknowledge that if different agencies were sharing more information, maybe we could have caught fraudulent behaviors sooner,” he said.

And as with other government issues, workforce management is a challenge. Ariga said the younger generation of “digital natives” recognizes the importance of data science, as does agency leadership because data science ties into the organization’s broader mission. But middle managers often must be convinced.

Try This

Ariga offered two ways for agencies to begin realizing his vision for data-driven decision-making:

  • Foster workforce talent. Government needs to accelerate its hiring process, especially because agencies can’t compete with the private sector regarding salary. Job descriptions need to accurately reflect the position (e.g., IT is not the same as analytics). And agencies should encourage professional development.
  • Modernize according to your data needs. When considering new technology, don’t simply swap one system or data center for another. Instead, think about opportunities to improve how your agency collects, synthesizes and then analyzes information. Is there a better pipeline to support data analysis?

This article appears in our guide, “Agency of the Future: How New Possibilities are Emerging in the Present.” To read more about how agencies are anticipating future needs, download it here.

Enregistrer un commentaire

0 Commentaires