The turn to spatial history has been aided by the explosion of digital mapping tools. While there are many options for mapping out there (including HistoryPin as described by Aaron Cowan in a History@Work post earlier this year), one look at the projects being completed by leaders in the field like the Stanford Visualization Lab is both inspiring and terrifying. How did they do that? Could I do that?
If you’re me, the answer is “not yet” (and not without a team and funding). But I’m increasingly interested in learning to make maps as part of my professional and scholarly work and wanted to stretch my digital muscles in some new ways. I just needed some data and a story that would be best told through a map.
Luckily for me, I found exactly what I was looking for in an archival folder that included a list of addresses where a community theater performed a particular play in Baltimore in 1980. I was embarked on a much larger project on cultural representations of Baltimore, and at first I didn’t see the value of this list of places. So I just took a photo of one page as a sample while I was in the archives and left it at that. Looking over my notes months later, though, it occurred to me that mapping these locations and overlaying them with census data from 1980 would give a good indication of who saw the show. I already had evidence from the director that people didn’t travel far to attend performances, which is why they did so many of them all over the city. With the map and census data, I’d be able to draw some conclusions about audience composition.
Easy, right? Get the census data and then plot some points on it.
Nope. Not easy at all. At least not for me. Someone versed in ArcGIS could probably accomplish this between morning coffee and lunch, but I’m a neophyte. And so I went to Twitter for help.
And got this helpful advice from Sharon Leon, good friend and guru at George Mason University’s Center for History and New Media. I started playing around with Google Map Engine Lite, and, while it’s limited in functionality in the free version, it offered an intuitive, flexible starting point that used spreadsheets, either uploaded or stored in Google Drive, for the data.
I began by putting all of the addresses into a spreadsheet. This is where knowing why you want to make a map is important. As with writing an article or essay, having a focus and a story to tell makes it possible to discard irrelevant information.
What I wanted was a visual representation of where the performances took place in the city, so I included the organizations and addresses. As I got into the data, I realized that the dates of performances could be interesting, too, and added that to my spreadsheet. After clicking “import” and navigating to my file, Google Map Engine asked which of my columns were the ones that pinpointed locations. After choosing address, city, and zip code, it asked me what columns I wanted as a way to identify the point on the map. I chose “name.” As if reading my mind, it assured me that the rest of the information in the spreadsheet would be transferred, too. In the end, all of it appeared in the box that pops up when each point is clicked.
In seconds, my map was made, though I got an error message about one cell in my table with an ambiguous address. This took some cross referencing, but I fixed it. Great. But looking at it now, I realized that my data included not only places, but time—those performance dates. Could I represent the “spread” of these performances around the city by month? Yes, by using layers. Each layer adds information atop the base map that can be toggled off and on by the user. The free version only allows 3 layers, so I made new spreadsheets with performance dates grouped in two month intervals. I uploaded again, and now one could see time and space.
I now had a basic map showing where the 1980 play had been performed around Baltimore. But what about the census data? That was much tougher. Charlene Mires, my colleague at Rutgers-Camden, clued me into Social Explorer, a database of historical census data that we have access to through the Rutgers University library. While I could download census data, the tables were too massive for Google Map Engine Lite to handle. Plus, in order for the map to plot the census tracts, I would need to upload shapefiles (which describe points, lines, and polygons that allow two-dimensional shapes to be drawn on a map) from the 1980 census (and anyone who can explain a shapefile better, please see me in the comments). I had no idea how to find those and, if I found them, if I could use them or how to with Google Map Engine Lite.
Rather than giving up or trying to find a new tool, I went back to my spreadsheets. While ideally I would have liked to be able to see demographic information as layers, I realized that nearly as good would be to have some specific information from the census in those hover boxes. Using Social Explorer, I navigated to the 1980 census and plugged in each address. Mostly I wanted race and class information, so I added columns to my spreadsheet for percentage white, percentage black, and percentage below poverty in each tract where an address appeared. Rough, sure, but a decent shorthand for my purposes. And while it was time consuming, it was easier than trying to figure out shapefiles, to be sure.
Check out the end result here.
In the end, my simple mapping project took me a few hours spread over a couple of days to figure out and is a useful reference tool for me as I continue my research. By making my data visual, it becomes obvious how the performances clustered in certain areas—and where there are outliers. Why those clusters and outliers are where they are is the job of the historian (hey, that’s me!) to figure out. Including the census data helps to remind me how segregated Baltimore was in 1980, with some performances taking place in tracts that are 99% white or black. That they occurred in both kinds of locales suggests the organizers’ understanding of these dynamics but also that mixed race groups were probably not seeing this show together. While it’s no Mapping Dubois—The Philadelphia Negro or the Orbis Network Model of the Roman World, my map has the advantage of being simple enough that I could do it without teams of coders or grant funding. Going through this process helped me learn the basics of mapping—and whetted my appetite for more sophisticated possibilities.