Solving Problems and Improving Performance through Data Analytics

When Superstorm Sandy hit New York City, Robert King was working in the Department of Homeland Security’s (DHS) Chief Readiness Support Office. During and after the storm, King and his office had to assess the damage to their assets across a wide urban geography that was experiencing a tumultuous natural disaster. This made taking stock of resources difficult, to say the least.

“There was a lot of calling,” he explained in a recent Viewcast with Grant Thornton Public Sector and federal analytics experts.

Lately, DHS has focused on consolidating data from disparate source systems within the organization, making it easier to provide integrated snapshots of information across the enterprise, especially at times when the information’s needed most, King said.

U.S. Government Accountability Office Director Vijay D’Souza also noticed the benefits of uniting datasets in a very basic but informative way when GAO assessed the federal government’s risk to earthquakes by merging FEMA hazard maps, federal property data from the General Services Administration, and workforce data from the Office of Personnel Management.

“This was static, but what if it was updated?” he mused.

Sharing the “what if?” attitude was Dominic Sales, Deputy Associate Administrator at GSA. Sales said that there were “1,000 flowers blooming around the agency” in terms of analytics and that he’s especially hopeful about the growth of predictive capabilities, artificial intelligence and machine learning. Though he acknowledges growth will be incremental, Sales is confident in the future.

“We are limited by our own imagination at this point,” he said.

But perhaps the magnitude of living up to one’s imagination—or just the idea of data preparation—feels intimidating. Shiva Verma, Principal and Decision Analytics Lead at Grant Thornton Public Sector, said that worrying about errors in data is a normal concern, especially considering the amount of data that organizations like DHS, GAO and GSA have to comb through and organize in order to provide reliable analytical models. But it doesn’t mean agencies have to stand still.

Data hygiene is important but shouldn’t unreasonably limit you, Verma said. Too often leaders are caught up in having completely perfect datasets.

“You can never have perfect data,” Verma said. And waiting for that “perfect data” comes at a cost.

The cost is inaction. The best move for an agency with data to use and problems to solve is to connect with data scientists who can help them move forward.

When setting up models, Varma said he has found success with a “center of excellence” approach, through which agencies build certain tools on a central basis with the enterprise in mind. When complete, those solutions are released for consumption while the center for excellence works on the “next big thing.”

That way, “there’s somebody always thinking about what’s next,” Verma said. “At the same time, you’re not getting clogged answering day-to-day questions.”

At GAO, D’Souza has had a chance to observe analytics at varying levels of maturity. He and his team have done everything from assess the usability of raw data in an agency to evaluate sophisticated analytic models generated in-house.

Of the adoption process, D’Souza said he believes that it relies on getting employees to constantly turn to analytics as a possible solution when they encounter an issue.

“The idea is not to make everybody in the agency a data analytics expert but to get people to think, hey, maybe I can use data analytics to solve a problem,” D’Souza said.

Since dealing with the chaotic Sandy situation and its aftermath, King said he’s seen an uptick in DHS’s analytic capabilities, which he believes will bode well for similarly stressful, decision-dependent scenarios in the future.

“Analytics changes the narrative,” he said. “Instead of asking for info, we’re immediately assessing the information.”

And those assessments will help leaders make more informed decisions throughout their organizations, Verma said.

“That’s when things get really interesting,” he said. “You start getting a return on your investment in all the data you’ve collected, the investment you’ve made in people and the time you’ve spent creating this framework.”

For more information about decision analytics in the federal government, watch the full Viewcast here.

This content is made possible by our sponsor. The editorial staff of Government Executive was not involved in its preparation.