Where are the geographers?

Do you know a geographer?

I’m a geographer and as I’ve navigated an evolving landscape of tech, data, and experience I have learned that almost everyone’s view of geography stems from 7th-grade Geography class. Some more interested/engaging people ask if I’m helping discover new lands, or if I work for the USGS, NGA, or Google Maps.

Geography is somehow such an abstract idea that is prohibits the discussion of geography’s roll in today’s world. Federal redistricting and gerry-mandering is a huge discussion now and all I see are geometry nerds and pundits talking — where are the geographers?

The first law of geography (which is surprisingly unrealized in many geospatial and GIS shops):

“Everything is related to everything else, but near things are more related than distant things.” — Tobler

This law is interesting because very rarely do we practice this type of awareness while working with geospatial data, or in a GIS, or web-mapping. I’ll elaborate: the New York Times loves popping out fast and dirty maps for sensationalized articles that get clicks — the only requirement is that the maps are “pretty.” I’ve had too many jobs where that’s my only requirement as a geographer. Pretty maps. Now in a world where clicks are easy and money is not, pretty maps win, and also (can) destroy the discussion. I believe pretty maps are great communicators because they draw attention, but that should be the final discussion before finalizing the map, not the only requirement. I was mortified by the Two Americas of 2016 maps popped out after the election. Such a broad, uninteresting stroke to draw and then sensationalize, but I guess that’s how data, maps, and news work together to undermine the conversation these days. And the liberal media (my media, by the way) is not helping to un-polarize our discussions by masking crude data as understanding, or heaven-forbid, truth. Geography should be a tool to learn and comprehend, not harshly categorize and divide. That’s Trump’s job, apparently. The reality is much more intricate, detailed, and caveated. Ocean and land to not have a hard boundary — there is transition, and in that transition space lies the secrets of change. Think of how we measure sea-level rise — we don’t measure it in Iowa, or in the middle of the Atlantic, though we’re getting better at that and those measures are important; we measure it where it has evolution, where it’s changing, where we see it — the littoral zone, the transition zone, the interface between one and another. In this vein, I dislike almost all election maps.

Another severely lacking geographic concept is understanding basic statistical principals within which confidence can be gleaned from a dataset. Broken down: we don’t see statistically significant geospatial analyses because no one understands the underling needs of such an analysis. I’ll start by dropping this tid-bit: this is why we have a Census. In order to state true statistical significance we must have a true measure of the Unit of Measure… so, if we are measuring qualities about human populations, we need to capture qualities of the domain of the population before we start sampling the population to determine statistically-significant survey/respondent results. This approach can be applied to any Unit of Measure; want to realize the spatial-autocorrelative effects of access to healthcare through locations and services in all health centers? — you must have the complete location and services data. If you want to learn about available health services and statistically-significant measures resulting from geographic proximity — then you must have every health center and its location before collecting survey responses. There are a few caveats to building geospatially-enabled statistically-significant analyses, but that’s a different post all together (a hacked together process can be gleaned from USAID’s Demographic and Health Survey’s process or from University of Southampton’s Dana Thomson’s grid-normalized survey process: my favorite).

I interviewed at USAID’s GeoCenter in 2014 and the Director and Project Manger didn’t understand why everything couldn’t have some measure of significance — with regard to 3rd-world data and mapping needs (data that will never have the label of “Complete” within a reasonable timeframe). The question threw me so far off that I realized I was talking to bureaucrats who had never read a book on geography… yet were in charge of using geography to help the poorest people of the world and millions of dollars of Federal investment. I gave them a shit answer and tried to get out of the Ronald Regan building ASAP, even purposely botching a question about my love for using R and Python — I said “Nah, I’m an Excel kid.” *shudder*

P.S. I ❤ R and Python.

I will continue my quest to apply geography in tech, and I hope posts like this might help with a discussion that I think is lacking.

Thank you for reading.