The federal government is vast, and the challenge of understanding its oceans of data grows daily. Rather than hiring thousands of new experts, agencies are moving to train existing employees on how to handle the new frontier.
Data literacy is now a common buzzword, spurred by the publication of the Federal Data Strategy 2020 Action Plan last year and the growing empowerment of chief data officers in the government. The document outlines a multiyear, holistic approach to government information that includes building a culture that values data, encouraging strong management and protection and promoting its efficient and appropriate use.
“While the Federal government leads globally in many instances in developing and providing data about the United States and the world, it lacks a robust, integrated approach to using data to deliver on mission, serve the public and steward resources,” the plan notes.
A key pillar of the plan is to “identify opportunities to increase staff data skills,” and it directs all federal agencies to undertake a gap analysis of skills to see where the weaknesses and needs lie.
It’s not the first time the government has addressed the issue of data literacy. In 2019, The federal CIO Council launched the Federal Cyber Reskilling Academy and received 2,300 applications. It ultimately chose 25 employees without IT backgrounds and trained them in basic cybersecurity defense techniques. The workers graduated but then found they couldn’t ascend the government’s pay ladder to new positions that required the skills. The problem was bureaucratic red tape that requires a mandatory minimum of time in a lower-level job before moving up.
“It was a beautiful experiment,” Dorothy Aronson, CIO of the National Science Foundation, told Fifth Domain at the time. “At the end, could they get hired? No. Okay, so what did we learn from that? … There’s a pool of people who are talented and could do it. Yes, you can train them quickly.”
EXPLORE: Which skills are most in demand in federal IT?
What Is Data Literacy?
Kristin Saling is clear on the definition of data literacy: “It’s whether someone can speak, read and write data and translate it into actionable information.”
As acting director of people analytics for the Army, she has her work cut out for her. The goal is to take 1.4 million service members and get them to a base level of data fluency, creating an ecosystem to fully data-enable the military branch. “Everyone needs to be able to use the cutting-edge data analytics platforms,” she says.
To better approach the vast challenge, the Army broke its personnel down into five groups: core data scientists, solutions implementers, leadership, data stewards and mission enablers.
That last cadre includes contracting and legal teams with little knowledge of data management. The service customized the curricula for each group, Saling says. “We want to make it as accessible and self-driven as possible,” she notes.
Ninety percent of the instruction will be in-house going forward. “We have a lot of skill and talent in this area,” Saling explains. That includes U.S. Military Academy junior faculty already posted throughout the Army. Notably, the training effort has strong support from the top brass. “We’re not taking time to convince people you need data in your life,” she says.
The Air Force launched a data governance boot camp pilot in 2020. The three-day training included synchronous online learning with the final challenge of each student taking the Public Sector Data Governance Test. All 100 participants passed.
“People are my most important resource,” says Eileen Vidrine, chief data officer for the Air Force. “Investment in them is critical to moving forward and an opportunity to optimize performance.”
The Federal Data Science Training Program
The Office of Management and Budget launched the federal Data Science Training Program in the fall of 2020. It included 61 federal employees from 20 agencies. The goal this time was to retrain workers to use new skills in their current positions. Key areas of focus included Python and R languages, machine learning, design thinking, data mining, data visualization, statistics, and enhanced presentation skills.
Students first completed 28 online courses delivered through LinkedIn Learning in the fundamentals of data science tools and techniques over 18 weeks. The second phase transitioned to virtual, instructor-led classes completed over 10 weeks. Participants completed more than 3,700 hours of online learning. Finally, students were required to deliver a final capstone project to their agencies’ leadership.
“Feedback from agencies has been overwhelmingly positive regarding the pilot cohort,” an OMB spokesperson tells FedTech. “The graduates are returning to their regular duties with a deeper understanding of and competencies with data science tools and techniques. We are carefully considering the outcomes of the program and determining how best to build on this success in the future.”
Leading the charge on data literacy are federal chief data officers, like Ted Kaouk at the Agriculture Department. The title is a new one at agencies and underlines the recognition that data is marbled through government organizations and increasingly drives missions. Kaouk says he prefers the term “data acumen” to describe the training effort.
The USDA is currently conducting a gap analysis per the Federal Data Strategy. “We’ve been assessing data skills broadly — visualization, quality predictive analytics, data prep and aggregation,” he says. “We need employees to leverage these skills to do higher-level activities.”
EXPLORE: Best practices for creating an agency data dashboard.
The USDA has ongoing training programs and runs data visualization competitions. Many employees have been recognized by senior leadership for their work.
“We have an increasing focus on data stewardship and quality, and that information is catalogued and accessible for decision-making,” Kaouk says. “We want to be focused on understanding our customers.”
Kaouk is also the leader of the federal CDO Council, which recently released a report on upskilling pilot programs at 10 agencies.
“Part of the challenge is that this is all holistic. As we see new needs, that creates new gaps. Because expectations are raised, that creates new gaps,” he explains. “We need the skill sets to do that work.”
DIVE DEEPER: Learn how artificial intelligence tools allow agencies to analyze vast amounts of data.
What Skills and Technologies Help Improve Data Literacy?
Federal agencies are using a combination of in-house and vendor training. Among the most common commercial educational platforms are LinkedIn Learning, Skillsoft, Pluralsight, Udacity, Coursera and MIT’s edX.
Agencies are also customizing teaching environments to their own needs and agency-specific data. “We are looking at the number of platforms that the Army already employs and at what competencies for data literacy they cover, and then both figuring out how to scale these appropriately and create savings and efficiencies, where possible, and figuring out where the gaps are where we need to create our own internal curriculum,” says the Army’s Saling.
When the Government Accountability Office announced its data training pilot last summer, twice as many people signed up as there were spaces. GAO Chief Data Scientist Taka Ariga says part of the idea was to gauge employee interest. With that question answered, the agency chose 25 workers who were analysts, specialists or operational professionals. At that point, the coronavirus pandemic was in full swing, and GAO used Zoom to deliver six hours of training.
“Our goal has always been to train our workers to be general contractors so they can supervise data projects,” Ariga says. “You need to know what right and wrong look like.”
Ariga says he was satisfied with the outcome, but some challenges arose. First, the instructors needed to be more familiar with the agency and its work. Second, middle managers should be approached to win support for a training program. “We need to speak data literacy to them in a way that doesn’t impact, time, budget and resources,” Ariga says.
MORE FROM FEDTECH: Which skills are most in demand in federal IT?
The Benefits of a Data Literate Federal Workforce
The Federal Data Strategy outlines the intended benefits of a data literate workforce in its guiding principles: responsiveness, anticipating future usefulness of information, ensuring relevance and harnessing existing data. All are aimed at creating a government that uses information to better do its job. For the military, the need to better handle information can be a matter of life or death.
“All DOD leaders must treat data as a weapons system,” said David Pearson, center director for engineering and technology at the Defense Acquisition University in a MeriTalk webinar.
The Department of Health and Human Services launched its Data Science CoLab in 2017 to boost basic and intermediate data skills. The collaborative program is the first try at a far reaching and cohort-based data-skills training for the agency. In addition to data analytics skills, HHS is currently training hundreds of employees on how to write Python and R.
“Demand for a seat in the Data Science CoLab has grown approximately 800 percent in the past three years, a testament to its success,” says Bishen Singh, a senior adviser in the Office of the Assistant Secretary for Health. “Beyond skill growth, it has led to incredible time and cost savings, as well as internal career growth for past participants across the department.”
The National Science Foundation was less successful with its Data Science and Data Certification Pilot, which had a class of 10 participants from various federal agencies. The workers were trained in advanced analytics techniques, with a focus on applying data tools to uncover meaning and solve Big Data challenges. However, the vendor curriculum used general data sets rather than agency-specific ones.
“As a result, participants found it more difficult to apply their learnings directly to real-world scenarios,” notes the CDO Council’s “Data Skill Training Program: Case Studies” report. The learning modules were mostly virtual and self-paced. Communication was poor with the vendor, and employees began to lag in completing their coursework. The pilot was discontinued.
Most of the training pilot programs were launched as the pandemic closed down government offices. The shift to virtual learning made progress difficult for some students. Another key lesson: Allow workers to use their new skills quickly, while they’re fresh.
RELATED: New data analysis tools help agencies move fast amid crises.