The FAA is awash in data. It comes from handling more than 45,000 flights a day. It comes from other sources like drones and other unpiloted vehicles. And it comes from other areas in the national airspace system.
The administration predicts the number of passenger flights will increase by 4.7% annually over the next two decades.
The FAA’s clearinghouse site for publicly available data is the home for much of this information. And there are 3,600 data sets on Data.gov. And this doesn’t account for all the internal data the FAA collects and shares with its employees.
This is why the FAA must apply new and emerging technologies to understand and make its data more valuable.
Marseta Dill, the deputy chief data officer for the FAA, said its data strategy is ensuring the agency is using data as an enterprise to drive insights, impacts and innovation.
“We manage a couple of data platforms and they’re based in the cloud. The goal there is that we want to provide them as a shared service. We’ve been focused on improving access to our data by onboarding those high-value data assets on to the enterprise repositories,” Dill said on the discussion Government Modernization Unleashed: Data and AI. “When our analytical community has needs, they can work closely with the CDO office to gain access to that data. It gives us an opportunity to deal with the data sharing part versus having the data stewards engage directly with a number of folks across the agency looking for access to data. We will can gain a lot of efficiencies that way.”
Workforce needs data tools
Those efficiencies are important because the FAA serves a number of data personas, including business analysts, data analysts, data scientists, data engineers and many others. Additionally, the administration provides these personas a plethora of analysis and visualization tools.
Dill said each of these tools serve a business need based on what the user community requests.
“I think as long as you’re engaging with your workforce about what their needs are, you can acquire the tools that serve a purpose. Each of these tools have different capabilities, some of them overlap and some of them don’t,” she said. “So although it may give the impression that there’s a lot of tools in the portfolio, they each have a unique purpose and function. Our goal is to help our users make informed decisions about the best tools to fit their needs.”
She said the FAA’s product roadmap is another way it drives efficiency as part of its governance efforts.
The roadmap is similar to an enterprise architecture that defines the current state of technology and the future state of where an organization wants to go.
Dill said the roadmap helps users understand what data tools and services are available today and what is coming in the future.
“I think there are great opportunities both in the private and public sector when it comes to artificial intelligence, especially when you think about access to big data. There’s lots of opportunities to automate tasks related to data, and it gives us a chance to really look at like some of the voids around the preparation aspect of data. How do we classify data? How do we look for anomalies?” she said. “I think machine learning and natural language processing gives us more capabilities to take full advantage of the data that we have access to, including historical data. When we think about our decision support tools, how do we identify latent risk? How do we make recommendations based on data that’s come in from multiple data sources? And, ultimately, we want to also think about ways to improve the way we engage in knowledge sharing.”
FAA leans into chatbots, RPA
One initial use of AI/ML is through a chatbot to help with community engagement.
Dill said the FAA typically receives about 300 inquires a month and many of them now can receive immediate responses instead of having to wait as much as 15 days.
“We work closely with our counterparts across the agency when they’re deploying applications like this. We have a strong partnership with our robotics process automation (RPA) team, and we recently just ran a campaign that was focused on leveraging not only the tools that we offer within the chief data office, but also through our RPA program office,” she said. “Part of what we’ve accomplished with these campaigns is that people can have a broader view of the tools that are available as they’re dealing with their unique challenges.”
As the workforce becomes more comfortable with chatbots, RPA and other tools, Dill said the goal is to make data and these tools more self-service. Employees will not have to rely on CDO staff or other experts across the agency and instead use forms or automated processes to register and use data assets.
“We put a lot of emphasis on establishing data governance communities. These are cross-organizational groups that are composed of not only data stewards, but also data users,” she said. “We’ve been able to leverage those communities to have conversations around what data is being leveraged in ensuring that that data is cataloged so that more people can be aware of the data that’s out there. With our data governance center, we’ve been able to automate a lot of the access to data assets. Those efficiencies that are gained can provide that service through our data governance center.”