5 Q’s for Brandon Pustejovsky, Chief Data Officer at USAID
The Center for Data Innovation spoke with Brandon Pustejovsky, chief data officer at the U.S. Agency for International Development (USAID). Pustejovsky discussed USAID’s efforts to have all of its partners abide by an open data policy and offered advice to other government agencies that want to be more data-driven.
This interview has been lightly edited.
Joshua New: USAID announced its policy for open development data in October 2014. Can you explain some of the key components of the policy, like the Development Data Library?
Brandon Pustejovsky: USAID’s open data policy establishes four key data management pillars. First, it defines USAID’s data governance structure by appointing a group of USAID executives to something known as the Information Governance Committee (InfoGov). This body oversees policy implementation worldwide, including the establishment of best practices for use by USAID operating units and implementing partners, compliance with U.S. government mandates, and ensuring the proper protection of USAID’s data assets.
Second, it requires USAID implementing partners to submit data generated with USAID funding to the Development Data Library (DDL). The DDL is the agency’s repository of USAID-funded, machine-readable data created or collected by the agency and its implementing partners. We envision the DDL becoming a major source of global development data over the coming years and a key tool which we can all use to make evidence-based decisions in support of our shared goal of ending extreme poverty.
Third, it outlines a data clearance process within USAID itself. While we rely to some extent on partners, Institutional Review Boards, and those closest to the ground to protect data that they know might jeopardize privacy or security, we also take our data protection responsibility within the agency very seriously. Once a partner submits data to USAID, it undergoes a rigorous clearance process by which our privacy, operations, legal, and data experts conduct additional reviews to ensure that proper safeguards have been taken prior to publicly releasing the data, including to determine whether the data can be released publicly at all.
Finally, it establishes “data stewards” in every USAID operating unit, including our worldwide missions, to provide guidance on the policy, to help mainstream best practices, and to assist with clearing datasets for public release.
New: Has there been any trouble getting every USAID partner to comply with the policy? If not, how did you ensure that the implementation didn’t encounter any roadblocks?
Pustejovsky: It’s been exciting to see datasets flow into the DDL as a result of the policy, and the overwhelming response among implementing partners has been quite positive. I don’t think anyone doubts the enormous potential that open data holds for improving development outcomes. However, I think we all recognize that global data sharing is a monumental shift but one that requires an artful balance between transparency and ethical data protection. We’re committed to finding that balance and to doing so via an iterative process that puts feedback from our partners at the center of our efforts. That’s why we’ve engaged with over 180 organizations on the policy thus far, seeking feedback at every turn and inviting additional thoughts via our presence on the Q&A site StackExchange. We also established a “Frequently Asked Questions” site last month, which has certainly helped partners comply responsibly with these new requirements.
You don’t reap the benefits of 20/20 hindsight until you take the first step forward. So we took that step, given the limited window of opportunity we had available, knowing full well we would need to make adjustments along the way. So we view this as a key year for learning and adapting with our partners to the new world of open data. This learning process is endemic to the way we do business. We know this is big. We know these changes aren’t going to happen overnight. It’s going to take time for all of us to adjust and refine our data management practices. But if there’s one thing I want our partners to know, it’s that every comment, every question, every recommendation is being addressed in one form or another. We’ve always anticipated an update to the policy based on this feedback, and we still do. We also look forward to providing more venues where that dialogue can take place.
New: Can you describe any success stories made possible with open USAID data?
Pustejovsky: During the 2011-12 drought in the Horn of Africa, USAID was able to reposition food supplies based on maps and visualizations created from open data, making our response more nimble and targeted to the areas of greatest need. It’s a powerful thing when you can start pairing your resource allocations with the situation on the ground, especially when you can make those decisions based on extremely current, or even real-time data. In 2013, M-Farm, an organization in Kenya, created a mobile application using publicly available crop price data to connect farmers with markets and to help them select crop varieties likely to yield the greatest income. We’re also looking at ways we can leverage open data, including satellite imagery, to monitor the progress of our projects under the Power Africa initiative, our effort to increase access to electricity for Sub-Saharan Africa. The potential here is just enormous, and we’re just starting to scratch the surface.
New: USAID has a far-reaching international presence, yet your team only has three employees. How do you ensure that the entire agency adheres to good open data principles?
Pustejovsky: These three employees are part of a much broader global team. This team includes over 100 data stewards in our worldwide missions and Washington operating units. These individuals are charged with socializing this new mandate and with guiding their teams on sound data management practices. At the same time, we are developing a compendium of references for our data stewards to draw upon as they lead data-driven efforts within their operating unit. These include best practices on data protection, the creation of codebooks and data dictionaries, data management plans, and standard operating procedures. In February, we also released an online, self-guided training focusing on USAID’s new open data policy, the benefits of open data, and where to find additional resources. We had over 100 individuals register for the course by the end of the first month.
New: As more federal agencies begin to recognize the importance of having a chief data officer, is there any advice you could offer, such as how the government can be more data-driven or better deliver open data?
Pustejovsky: USAID has long been a data-driven agency, and I think that holds true for much of the U.S. government. But the reality is that our data—the taxpayers’ data—has been stored according to various standards and formats that have come into being during the global sprint into the digital age. The challenge now is finding a way to relate these data assets to each other, to yield a more comprehensive view of taxpayer investments, even if the underlying data structures are different. You can’t simply assume that the term “district” in one dataset means the same thing as it does in another dataset, for example. This is a monumentally complex task but one that’s absolutely worth doing. This dynamic plays out within individual agencies, as it does across the U.S. government when separate agencies want to share data with each other. At the same time, everyone is working on this aggressively. So I don’t think there is a lack of political will to become more data-driven. If anything, the political will is there in spades. If the pace of change signals anything, it’s the enormity of the task of getting data to “speak with one voice” at the intra- and inter-agency levels. This is by no means insurmountable, but it does take time.
Regardless of whether or not an organization has a chief data officer, I think the best thing you can do is listen to your stakeholders and be proactive in responding to feedback so that you can deliver data more effectively. USAID’s mission statement begins with the phrase, “We partner.” Our emphasis on partnership stems in part from the fact that we recognize that we don’t have all the answers. We need to collaborate, learn, and adapt if we are to improve development outcomes. And we simply cannot do this without listening to the needs of our partners and other global thought leaders. And as you’re making these changes, you need to communicate along the way. That’s one of the reasons we created our FAQ site and why we do these interviews. Have you never noticed that the best companies don’t just leave you on hold while they work on your request? They check back from time to time to let you know what progress they’re making. That’s the kind of organization we want to be. We’re listening; we’re iterating; we’re moving forward. There is still so much left to do, but these are exciting times and the best days are ahead of us.