GSA data skills catalog promotes ‘learning culture’ of Federal Data Strategy

In Featured by ThePeopleLeave a Comment

Despite a tumultuous year, the interagency team behind the Federal Data Strategy made significant progress increasing data proficiency across the federal workforce.

As part of the strategy’s 2020 Action Plan, the General Services Administration has developed a Curated Data Skills Catalog and a Data Ethics Framework to help agencies develop competencies when acquiring, managing, and using data to support their agency’s mission.

Trey Bradley, a program manager for strategic data initiatives at GSA and member of the Federal Data Strategy team, said GSA’s data skills catalog is meant to develop a “learning culture” among agencies meant to develop data leaders at all levels of proficiency.

The Chief Data Officers Council, which held its inaugural meeting a year ago, is responsible for maintaining and updating the skills catalog, and shares responsibility with the Interagency Committee on Standards Policy (ICSP) for updating the ethics framework.

Bradley said the CDO Council is “already engaged and thinking about next steps and enhancements with both of these efforts. Nothing firm has been promised yet, but I know they are thinking about case studies, use cases, playbooks, etc.”

The skills catalog and the framework satisfy two of the 20 interagency action items listed in last year’s action plan. The COVID-19 pandemic caused some delays, and pushed some deadlines back by up to three months. Despite these challenges, Bradley said GSA has made significant progress in completing its work.

“It was a challenging year, but agencies rose to the occasion, and all the specific action steps were completed or partially completed,” he said.

This work is expected to continue into the new year. A 2021 action plan remains in development, but Bradley said it will focus on the same themes as last year’s strategy.

The skills catalog highlights a range of data-skills training available to the federal workforce, from training non-technical managers and staff to understand the basics of data to statisticians and data scientists who continue to hone their skills. Bradley said some of this training is available to all federal employees, while other coursework serves as an example of what other agencies might do.

“I’m not going to say we’ve addressed the training challenges completely,” he said. “We need to make training more widely available and available to everybody — not just the more technical staff. We did identify some real pockets of excellence in our work, however, and believe the government has a strong foundation to build upon.”

The catalog, for example, highlights the work of the Census Bureau, which has stood up a reskilling program focused on improving data science skills at all levels of its workforce. This consists of online coursework, training from Census subject-matter experts, a mentorship component and a capstone project that gives participants a chance to put their new skills to the test using real bureau data.

The bureau’s first cohort completed the reskilling program last June. A second cohort will begin later this month and will offer two separate tracks for data generalists and machine learning.

The catalog includes a Federal Data Lifecycle, a framework that outlines the skills necessary for all stages of handling data, including collection, analysis and dissemination.

GSA, meanwhile, has released its ethics framework with the intention of helping agency employees and managers manage risk, support transparency and build public trust.

The framework touches on the steps agencies have taken to build an ethical foundation for using artificial intelligence and mitigate bias in these tools. The Defense Department and the intelligence community have developed AI ethics principles, and more recently President Donald Trump signed an executive order last month giving agencies a shared ethics framework for developing and using AI.

The ethics framework mentions AI and bias in one of its cases, which Bradley said boils down to being aware of bias, making every effort to mitigate against it and being transparent with algorithms and methodology.

“Most of this bias, I believe, is unintentional. That said, that doesn’t mean we don’t have the responsibility to try to prevent it as much as possible,” Bradley said.

Leave a Comment