The Center for Investigative Reporting’s collaboration with NPR on the Borderland multimedia project was an unexpected excuse to work on two of my passions: the U.S.-Mexico border fence and historical mapping.
Journalists at CIR have been trying for more than three years to obtain accurate, detailed mapping data showing the location of the border fence system. The fences – it’s more accurate to say there are many – have cost taxpayers many millions of dollars and are key pieces of border security infrastructure. They’re also a potent symbol that chafes at our Mexican neighbors.
We filed several Freedom of Information Act requests with U.S. Customs and Border Protection, and after several appeals, we received limited data showing where individual fence segments start and end. But we were told repeatedly that the actual lines showing the details of the fence segments were sensitive law enforcement information that could give away secrets to drug cartels, illegal border crossers or terrorists.
This reasoning always has bothered us: How secret is a 10-foot-tall metal fence that runs along golf courses and through major cities? Wouldn’t any smuggler worth his or her dope know more about the tiniest minutiae of the fence than border agents?
A positive twist on this frustration was the realization that because the fence usually is big and clearly visible, we could trace it using publicly available satellite photos. So, because I’m stubborn, I spent several weeks doing just that in 2012, panning through mile after mile of photos and trying to discern the difference between security fences and canals, roads and livestock fencing.
The mapping program I used, JOSM, allows us to digitally trace the lines of the fence in satellite photography that was likely (and ironically) originally pulled from public U.S. government sources. That tracing gave us a new data source that could be exported and used in other GIS (geographic information system) software.
The result: We now have what is – as far as we know – the most complete and detailed map of the border fence system that is publicly available. This allows us to show great details not visible in zoomed-out views of the fence. For example:
- Gaps between hills:
- Open gates on the Tohono O’odham Nation reservation in Arizona:
- And the fence winding past a golf course:
Is the map a 100 percent accurate, definitive document? No. My hope is still that at some point, federal officials will release a full dataset of the border fence system to provide an official record.
But when we (or you) find issues, the good news is that we (or you) can fix them. We chose to draw the fence in OpenStreetMap, a community-generated geodatabase of the entire world. This wasn’t an obvious choice. It meant that everything we did was done in the open – at any time, others could have discovered our work and used it in their own publication. We could have been scooped on our own reporting.
So why did we do it that way? Partly because this was a passion project – we didn’t have a firm idea of when the data would be used – and partly because CIR encourages sharing whenever possible. Even if we never got to use the data, we wanted to make sure the work wasn’t lost.
Working in OpenStreetMap also made it possible to play with the data in many formats, which would make sharing easier if and when a partner came along. That partner turned out to be NPR.
We connected because NPR was researching border data as its staff prepared to travel for the Borderland series. As it turned out, we had a lot to share.
In addition to the current fence map, we contributed a lot of research and experience about the development of the fence and helped generate a history of the border itself, from the time of Mexican independence from Spain in 1821 to the modern border’s completion in 1853.
I’m always amazed by what spatial data is available if you dig around enough. Due to the 20-year project edited by John Long for The Newberry independent research library in Chicago, you can get digital boundaries for U.S. counties going back hundreds of years. We used those files and other geospatial data to derive historical boundaries between the United States and Mexico. (In 1821, for example, Mexico included much of the land that we know as the American West.) That by itself was a really fun project for me and Matt Stiles at NPR.