The past two years the Agriculture Department spent making some major investments in its data infrastructure have paid off during the COVID-19 pandemic.
As part of this enterprise analytics modernization, USDA stood up more than 500 dashboards that cut across mission areas and C-suite offices. These dashboards have grown in popularity among agency leadership, logging about 40,000 views a month, but in recent months, under the pandemic, those views have more than doubled.
“COVID-19 has been a real game-changer in terms of the demand for data, the need for data and really highlighted some of our opportunities and challenges,” USDA Chief Data Officer Ted Kaouk said Thursday during a presentation hosted by ATARC. “We’re seeking to make sure that decision-makers have the data that they need and the public has the data it needs to make decisions about public safety and continuity of operations.”
As the chairman of the governmentwide CDO Council, Kaouk said his colleagues established a working group specifically aimed at answering data-specific questions about COVID-19.
USDA’s data dashboards have played a significant role in agency decisions to reopen facilities across the country. James Barham, the acting chief data officer and the director of data analytics for USDA’s Rural Development, said the demand for in-person agency services has been felt most acutely by customers who are unable to seek assistance online.
“In some cases, we need to be there. With our rural housing and our single-family housing that goes to very low-income communities, they’re not online. They make payments sometimes in-person — same thing with our farmers as well. So it was critical to get people back in the office, and of course, do it in a safe and healthy way,” Barham said.
USDA’s enterprise data infrastructure has also allowed Rural Development to stand up a COVID risk portfolio that shows the economically hardest-hit counties in which the agency can provide assistance in the form of loan deferrals.
In addition, Barham said Rural Develop has its long-term program evaluation and day-to-day program performance operations working together to track data trends during the pandemic. Rural Development operates as a major lender with more than $250 billion in assets.
“I’ve got a suite of economists that are doing this longer-term evaluation, impact analysis work, and then a suite of data scientists that are running the daily numbers on [loan] delinquencies and portfolio performance,” he said. “We’ve had to merge these two teams like never done before the pandemic for the need to be able to understand on the ground impacts in rural communities and how it affects our portfolio.”
While USDA’s data infrastructure has proven useful in its eight mission areas, Deputy CDO Chris Alvares said the enterprise analytics platform also allows agency leadership to get a big-picture view of how USDA is operating across all of its operations.
“We really had a need within USDA to bring our analytics capabilities together into a common environment, a shared resource that starts to break down some of the silos that have been barriers to accessing data, to leadership getting insights, to our workforce from all levels of the organization having access to the same pieces of data,” Alvares said.
As for next steps in this data journey, Kaouk said the agency will complete its workforce data skills assessment in the next week or two.
The assessment will not only determine where USDA stands in building a data analytics workforce, but also give the agency a big-picture look at where its entire 100,000 employees are standing in terms of data literacy skills.
Cynthia Parr, the acting assistant chief data officer for research, education and economics at USDA’s Agriculture Research Service, said the workforce assessment would also help determine where the agency stands in terms of skills needed for artificial intelligence.
“Part of the goal there is to figure out what is our baseline for the workforce. Not just the folks who are actually doing the AI, but everybody who is actually in that data pipeline — to make sure that they understand not necessarily how to do the AI, but how it will be used, how to do the best they can to make sure that the data going into those models is high-quality and that they understand how to interpret the results that come back out. It’s really a very holistic perspective,” Parr said.
Kaouk said USDA’s 500-plus dashboards have saved about 400,000 in labor hours that previously went toward manual tasks like checking the status of funds.
“We are really interested in the broader needs that these kinds of changes have on the skill sets that are needed,” Kaouk said. “As we reduce data calls, we make tools and data more available, what opportunities we have for staff to focus on higher-value activities such as analyzing data.”
USDA is also rolling out a data strategy through 2023 focused on four goals – data governance and leadership, data and analytics workforce, common data and analytics platform and open data.
“We feel it’s fairly important, now that we have good governance around the data that we have leaders in place to enable a strategic approach to our analytics development, integration of data and data-sharing more broadly,” Kaouk said.