The Center for Data Innovation spoke with Chris Rasmussen, public open source software development lead at the National Geospatial-Intelligence Agency, a federal agency that provides geospatial intelligence to support military and civilian operations. Rasmussen discussed what open data means to a defense agency as well as how the open source movement has been transformative for how the government approaches geospatial data.
“I’d estimate that if you factor in the total investment, there’s probably $10 million worth of code. To make that free is an enormous effort, but this shows the goodwill that the agency is investing in making this software available.”
Joshua New: Most people probably have not heard of the National Geospatial-Intelligence Agency (NGA), despite it being at the forefront of the U.S. government’s geospatial data efforts. Could you discuss the role of the agency and why geospatial data is so valuable?
Chris Rasmussen: The NGA was formed out of smaller intelligence and mapping agencies and it was designed to be a sum greater than its parts. The primary function of the agency is to combine geospatial and temporal data to provide context to intelligence issues. When you have combine two different organizations—a mapping culture and an imagery analysis culture that came from the Central Intelligence Agency—together, there are a couple different key aspects. Some on the more unclassified side, like the safety and navigation missions, have often been public. For example, if you look into Apple and Android app stores, one of the apps that we’ve released called ASAM (Antishipping Activity Messages) is designed to inform the public about piracy and display hostile acts on the seas to the unclassified world. And then, there’s a large portion of the organization that’s traditionally known as imagery analysis, which is what makes sense of data from satellites, airborne sensors, and other things. This part has been growing lately with the commercialization and commoditization of the discipline itself is the number of commercial sensors, rather than government ones, and the amount of volunteered geographic information and the amount of pictures people are taking on the ground. All these parts combine to provide the intelligence community with geospatial situational awareness for government purposes.
New: NGA’s data and technology is useful for a wide variety of civilian applications as well. Could you discuss some of these?
Rasmussen: GitHub and app stores us give us a way to push things to the public, so why not put certain applications to where the biggest audience is? There are government app stores but they’re pretty limited. When you’re trying to build a crowdsourced geospatial data repository, you obviously can’t do it without a crowd. As far as audience goes, Apple’s app store has a billion people in it—if you’re trying to get people to use your apps, that’s a pretty good place to put them. So our apps like ASAM and the Disconnected Interactive Content Explorer, which gives turn-by-turn navigation functionality when you’re disconnected from a network, which have important civilian uses, now can reach the audience that needs them.
The agency has a substantial safety and navigation mission, and a lot of this mission is on the civilian unclassified side. That’s been a function of the organization for many many years. Of course there’s a lot of classified missions, but in addition to safety and navigation, the agency also has a big unclassified domestic support and humanitarian disaster response mission. We’re still an intelligence agency, but we also have partial civilian oversight so we’re unique from an authority perspective—somewhat like the national guard. We may not have a mission to do what FEMA does, but they can tap us to provide help with geospatial data. Over the years, we’ve provided substantial support after hurricanes and other disasters in the United States and overseas.
New: NGA launched its Pathfinder project, which you head, earlier this year. What’s the goal of the project?
Rasmussen: The goal of Pathfinder is to invert the organization, particularly on the intelligence side. If you look at how geospatial intelligence has been commoditized and commercialized, particularly with smaller satellites, the amount of data from social media, and volunteered information, there is so much value on this unclassified side. Our base of operations, if you will, right now is classified, and we kind of sprinkle some open source on top. That needs to be completely inverted. Our default needs to be to go to the unclassified side first and then focus on the remaining gaps. Pathfinder is trying to instill a “learn by doing” approach to explore this. It’s going to help us figure out how policies, the trade craft, the research, the hiring practices, and so on will need to change. For example, we’re finding that you don’t really need clearances when so much of the data is on the unclassified side. This is an incredibly complex challenge, but Pathfinder should help convert the agency into a more open-source research organization.
New: Traditionally, open data has not generally been associated with the defense community. For example, U.S. open data policy exempts national security systems from open data publishing requirements. You mentioned NGA has a GitHub Page, but how else are is the agency dealing with the shift to open government data?
Rasmussen: From a consumption perspective, NGA and other defense agencies are increasingly consuming open data. We’ll pull from data.gov or any other platform we can pull unclassified data from. The issue is talking back, and figuring out how we put out more open data is a huge challenge. For open source tools, publishing data where the largest audience is just makes sense for utility purposes. We’ve tested some outputs on data.gov. For example, earlier this year we opened a lot of Ebola data to test how we could push this out in an operational capacity so more people can help analyze this data and figure out what’s out there. We tested it, and it does work, but it was more in the “research and development” bucket at data.gov. Pathfinder should be a huge help for us to ensure that more and more data can be open.
New: How does contributing to the open source movement relate to NGA’s mission?
Rasmussen: Simply put, we can’t succeed without it. Intelligence officers and analysts are knowledge workers, and the knowledge worker toolkit is growing with new open source tools and datasets to exploit. The reason that we started open-sourcing our tools on GitHub is because we have every interest in growing high-quality software, just like everybody else. To be a part of the “big data” conversation, you have to be in that ecosystem. It helps the agency improve the software it’s using. We have a new tool called Hootenanny, for example, that we want to run as smoothly as possible. This means getting all the bugs out, which means exposing it to the best talent possible. At this time, millions of people with this talent are on GitHub.
If you look at our GitHub page, we put out a ton of high quality software. I’d estimate that if you factor in the total investment, there’s probably $10 million worth of code. To make that free is an enormous effort, but this shows the goodwill that the agency is investing in making this software available. When we talk to organizations we need help with, we can say we’ve invested a lot of money in making the best possible software with the best help possible and that we want to be transparent about what we’re doing. This is what brings in new ideas—from the public and from companies—that never would have happened if we kept acting like a “spooky” intelligence agency when we didn’t have to.