How to Build the Data Skills Toolkit Public Employees Need
By Danuta Egle and Andrew J. Zahuranec
Last week, The GovLab and Center for Urban Science and Progress (CUSP) — both based at New York University’s Tandon School of Engineering — hosted a webinar in which participants discussed their vision for a data skills toolkit for public employees. Facilitated by Victoria Alsina, academic director at CUSP, and Stefaan Verhulst, co-founder and chief research development officer at The GovLab, the webinar explored the key skills and competencies of data scientists working in the public sector necessary to tackle critical and complex urban challenges.
Despite the critical application of data science across industries, the use and re-use of government data continues to pose challenges for public sector officials and agencies. One significant barrier to the application of data science in government decision-making is the evident lack of skilled data experts within the public sector.
This webinar asked panelists Ai Yamanaka (Port Authority of New York and New Jersey), Amen Ra Mashariki (World Resources Institute), Francesco Marconi (Applied XLabs), Laura Kahn (Accenture Federal Services), Lilian Coral (Knight Foundation), and Neil Kleiman (New York University) to jointly reflect on the following topics in aligning expertise and education:
- What strategies can public officials develop to ensure the risk of discrimination and bias does not increase with the use of data-driven methods?
- Does it make sense to structure training programs along the data life cycle?
- In addition to the technical data expertise, which are the other key governance and business skills and competences needed?
- How can we nurture bi-lingualism — where data skills are matched with a particular domain expertise?
Discussion Recap
Beyond Technical Expertise
During the discussion, several panelists emphasized the growing disconnect between data science and its applications. The issue, they noted, stems from data scientists’ training, which requires more than just technical expertise to meet community needs.
Lilian Coral explained that the integration and use of data in society is characterized by the three phases of data: opening up data; developing data analytics; and making data useful for the public. Each phase dictates a different type of training. The first and second phases required technical and analytical skills to collect, construct, and interpret data in a meaningful way. Today, however, the public sector needs a previously untapped skill: contextualizing data to reflect and respond to real-world problems.
Consequently, an ideal data training program would need to expand technical and analytical skills, show students how to communicate with data, and give them digital tools to contextualize data.
Neil Kleiman agreed. Throughout, he emphasized that universities taught more than just programming skills. They provided students the context needed to use them.
“If you talk about technical skills, people can get that anywhere,” said Kleiman. “What they can’t get is the context for when it’s appropriate to apply the data.”
Bridging the Gap Between Public Interest and Data
The skillsets of data scientists and domain-specific experts in the public sector vary. Unfortunately, several panelists stressed, the distinct training of public employees can lead to a language barrier when communicating with others. In other words, the public sector suffers from a supply-demand problem.
Ai Yamanaka framed this issue as one of communication. Developing a relationship between data scientists and domain-specific experts requires good listening.
Often, data scientists don’t have expertise in the area they work in, so they need to be able to defer to those who do. It also requires them to present data in a meaningful way for the intended audience. Visualizing data and crafting a narrative are key skills for data scientists in the public sector.
Amen Ra Mashariki noted this deference was connected to another need: accountability. While community members can understand the responsibilities of other public offices, many have difficulty conceptualizing the role of data officers in day-to-day urban problem-solving. To serve their communities, data officers and specialists need an understanding of the governance structures, community engagement, and operational leadership of local governments.
Learning how to engage the ecosystem of organizations in urban centers that support government work, Mashariki added, can ensure sustainable, progressive data science excellence.
Redefining a Data Science Curriculum
Toward the end of the discussion, Francesco Marconi emphasized that impact-centered training is crucial because data is not neutral. Indeed, the frequently cited risks of data-driven decision-making include the misrepresentation, inaccuracy, or hidden biases of data. A new data science curriculum should encourage finding multiple methods of data verification and collection.
The panelists also discussed how expanding data science curriculum should also inquire into the ethics of data and emerging technologies for its collection and archival. Laura Kahn anticipated an interdisciplinary approach to data skills training that will look into the accountability of officers and technologies, while also focusing on more complex urban challenges that integrate several domains.
Kahn agreed with other panelists. The new data science curriculum, she argued, should place a greater emphasis on communication, strategic thinking, and storytelling.
Although the existing data science curriculum does not gear students toward civil service work, many recent graduates seek to leverage their skills for positive impact. Marconi drew attention to a new generation of data scientists and specialists that demonstrate an increasing desire to serve the public interest.
Takeaways
Across the many topics discussed, the panelists focused on three priorities in designing a data skills toolkit for public employees. First, everyone agreed that training is not just technical. A new curriculum should examine the application, presentation, and use cases of data science in the public sector. Second, panelists emphasized the importance of communication between data scientists and domain-specific experts. A new data skills training for public employees should consider first identifying the relevant organizations and institutions of the urban ecosystem through a matrix of people engagement. Finally, an interdisciplinary, impact-centered training should incentivize public employees to critically examine and interrogate data for inaccuracy, invalidity, and bias.
This webinar brought together a range of public officers and experts to begin a conversation on the next generation of data scientists in the public sector. At The GovLab, we were excited to continue our collaboration with CUSP to identify target areas for a new curriculum for public employees. This effort stems from The GovLab’s ongoing work to foster a new generation of public entrepreneurs through training offerings like The GovLab Academy, Collective Crisis Intelligence course, and data collaboratives masterclass, as well as our work to define new institutional functions and responsibilities, such as data stewards.
