Matthew Bafford’s Musings

Mapping the Washington, DC Metro

This project is largely irrelevant, as the WMATA website now has a live Google Maps version of their rail map with the same exact information available. I did this long before they provided such a service and keep it around just because I found the project interesting when I did it.

Having access to GIS software well before Google Maps made public consumer mapping popular and prevalent, I got curious as to how closely the public maps provided by the Washington Metropolitan Area Transit Authority (WMATA) matched the physical layout of the tracks. Obviously the rail map provided by the service is fairly stylized. A full and clearer image of this map can be obtained by going to the WMATA Rail Map on their website. I've shrunk it here to hopefully fit under fair use and not get harassed for copyright infringement.

My first thought was that the outer stations seem to be a lot closer together than they are in actuality. I also wondered how much obfuscation was done for security reasons. The stations are all public knowledge, of course, but sometimes these things are hidden. The actual underground rail paths are still an unknown to me, of course, but I imagine those are public record as well. I started by parsing out the addresses from the Station Listing on the website. Each station is in its own page and so I just pulled down all of the pages and parsed out the addresses. This gave me a listing of station "colors" and addresses. Next I took the addresses and geocoded them using ESRI's ArcMap product. ArcMap is a professional level GIS software package that offers data processing, mapping, and geocoding support. These days, the same functionality as I used ArcMap for could easily be found in free services available on the internet. Yahoo, for example, offers a free geocoding API that can easily be used to convert these addresses into real world coordinates. Google Maps, or OpenStreetMap are both free mapping interfaces. Once I had geocoded all of the addresses I had a list of points that made up the station lines. With a little connecting (manually done, given the small number of points), I ended up with a vector map of the network. Finally, I overlayed the vector data I had just generated on some base map data of the area that I had available from another project at work.

 

And, for reference, the map provided by WMATA follows.

 

This pretty much confirmed my suspicions. The WMATA map takes a lot of liberty with the Orange line (those stops west of Rosslyn are pretty far apart), but overall does try to give an indication of how far apart two stations on the same line are. It's very difficult to produce an accurate map that is quickly legible and that also conveys the important information quickly. Most users of the metro system don't care how far apart two stations are, so not as much emphasis is placed on preserving this information. All in all, I think they did a pretty good job. I'm curious if the NYC subway preserves this information more effectively. Given the size of the railway railway transit system, the maps are larger and much more detailed, which allows for more accurate preservation of the true spatial information.