Food security is arguably a geographic problem. Where it’s grown or farmed, the roads it travels to market, the population densities and their relative position to the source of food: These are geospatial considerations.
With this in mind the National Geospatial-Intelligence Agency hosted a hack-a-thon in March in Los Angeles, one in a series of events in which volunteer coders gather to apply geospatial solutions to critical issues in national security.
“Food security is of interest in a range of areas, from national security to humanitarian crises,” said Air Force Capt. Adam Satterfield, NGA’s lead for crowd source-driven innovation. “It’s not just of interest to the Defense Department and to the NGA. The general public has an interest as well.”
NGA hosted its first hack-a-thon in May 2016 in Huntsville, Alabama, in an effort to address emerging first-responder capabilities. The agency has since been betting heavily on this citizen-driven innovation mechanism, hosting four more events last year and another disaster preparedness hack-a-thon earlier this year.
“They are changing the culture of the agency,” Satterfield said. “You are seeing a lot more people who are comfortable exposing our mission problems to a group of innovators. The partnerships we are creating with academia and industry are bringing great value to the agency, helping us to think about how we can use the newest tools in order to stay relevant and stay ahead of some of the problems we are facing.”
The Los Angeles event drew 56 participants, about 60 percent of them students from the University of Southern California and the rest drawn from across the general community. Planners gave the coders Morocco as a test case, supplying them with data from Descartes Labs, Planet Labs and other sources and tasking them to come up with creative solutions to local food-availability issues.
Prizes included a cash award of $3,000 for the winning team. The winning team considered ways to fine-tune the delivery of fish. It’s a big issue in Morocco, where 40 percent of fish spoils on its way to market.
“They looked at where fish are harvested and how they could optimize that distribution based on hunger,” Satterfield said.
Algorithms took into account port locations, road networks, population centers and other factors in order to develop an optimized solution for distribution.
The runner-up team considered location-based approaches to crop management. Coders looked at local staples like corn, rice and barley, correlating data on soil types, population densities and historical crop yields. Taken together, this information could help farmers get a read on where best to plant for optimal yields.
Another team developed a computer model to predict what areas are most likely to encounter food security issues. They used a normalized difference vegetation index, a tool that helps researchers to get an easy visual read on whether an area contains live vegetation. By combining this information with other geospatial factors — road availability, local crime rates and other geographic hazards — the model could help preempt food shortages and avert future humanitarian crises.
For NGA, the event offered not just point solutions to specific issues in a given locale, but also a means to tap into new and creative approaches to geospatial problems.
“One of the real benefits of these types of events is that it allows us to interact with people who can look at problems differently, who can correlate data points that we haven’t thought of before,” Satterfield said. “Even if we don’t get a refined solution, just the methodologies that emerge can be extremely beneficial.”
In the bigger picture, the hack-a-thons could serve as a long-term recruiting aid for an agency that does not have high public visibility.
“You can’t recruit talent if you are not out there advertising the mission set.” Satterfield said. “We have really relevant data sets and interesting problems, but you can’t recruit the best folks if you keep those things behind closed doors.”