Mapping the Midwest: Keynote at the Midwestern History Association
These are remarks I delivered during my keynote at the Midwestern History Association conference in Grand Rapids, Michigan, in June 2017.
I keep a stack of mapping reference books near my desk, and two in particular are important: The Historical Atlas of the New West and the Historical Atlas of the American West. I keep these handy not only because I am intensely interested in maps, but also because when I’m writing or teaching the history of the West I try to bring place into my work. The where of the past is just as key as the when.
But if you were to crack these books open, you’d notice that the Midwest is missing. If you get a glimpse of the Midwest, it’s likely to be fleeting: a transitory place notable for the Transcontinental Railroad or Bloody Kansas. But you won’t find historical atlases that explore urban places, the region’s religious history, its ethnic diversity, or its historical ecology.
Perhaps this isn’t that surprising. After all, these are atlases of the American West, not the American Midwest — perhaps it’s fine that the midwestern region isn’t included as part of the West.
But, to my knowledge, there’s no specific historical atlas of the American Midwest. There are specific focuses on states, such as Nebraska, or overlapping regions, like the Atlas of the Great Plains. But nothing that focuses on the Midwest as a region — that attempt to use maps to explore and explain something about its history, or that uses historical maps to understand how people perceived the region.
I’m here today to talk about these missing maps — this lost region. But I also want to speculate on what we should do next. I want to talk about why we need these maps; I want to argue that print maps are not the way to achieve this; and, I want to make an argument for a new historical atlas grounded in digital maps. All of this is to try and answer two questions: first, can digital mapmaking overcome the limitations of print? Second, if we were to build a digital Midwestern atlas, what are the next steps for us to consider if we build maps that are complex, nuanced, and layered in their use and understanding of the past?
Today, then, is part show-and-tell, part speculation.
There is a lot of data available to historians of the Midwest — we can use population data, we can look at religious denominations, we can look at transportation systems, at demographics. But with so much historical information available, we’re also faced with a set of limitations. And that limitation belongs to the world of books. After all, print maps can only convey a set amount of information: there’s only so much data they can contain before they become unreadable; and, we can only print so many pages in book before the expense becomes too high.
In other words, print maps make it difficult to convey the things that we as historians are interested in: change over time, and difficult to examine those changes over time at different geographic scales — region, city, neighborhood.
Interactive, digital maps provide a solution. We have a better way of using detailed sources, of changing the views very easily in our data, to build richer maps that can tell us not just more, but more humanistic, things about midwestern history. But we need digital maps to get there: to work at different geographic scales, to display change over time, to integrate maps and our sources, and to craft arguments and narrative.
This is often referred to as a deep map, a map that not only visualizes data, but permits access to sources underlying the data. Deep maps also embed narratives, guiding readers through the spatial dimensions of an analysis. Deep maps usually provide a variety of ways to look at a place.
There are two key problems with print maps, as Laurie Maffly-Kipp has argued in the context of religious atlases. First, our sources; second, definitions. She argues that data tends to be limited, that historians have tended to map the things that are easy to count and we already know a great deal about. Which leads to a problem of definition: by sticking so close to such data sources, we tend to limit a more complex and layered understanding of historical events. Print atlas as have devised clever ways to deal with complex questions when faced with the constraints of print. But print maps are still fixed in space and time. What if we could instead use digital interactive maps? How might we envision a midwestern atlas? How could we build more humanistic maps?
Part of the solution is to use animation and interactivity to identify broad trends. The map below looks at city population: the red circles represent the population of cities; the largest the circle, the higher the population. The shading indicates county population density: the darker the blue, the higher the density; the more green to yellow, the lower the density.
The above map is fairly interesting: letting us explore the broad population trends in the Midwest. By animating the map, we can see growth and decline, find cities whose populations shrink as well as those who grow. But we can do more.
Here we start the speculative section of our time tonight.
Where this map excels at is showing population trends: it’s the nature of our data, as Census records are tied to the county level and paired with the most complete set of city population data that I’m aware of. And these flows of people are big questions for the Midwest. Between 2000 and 2009, Midwestern population growth concentrated in and around metropolitan areas, while rural areas have continued to lose population. Not a single state in the Midwest has added to its rural populations in this period. Omaha ranks as the region’s 12th largest population; Sioux Falls had the fastest rate of population growth, growing by 27 percent. The Midwest’s pattern of metropolitan growth and rural decline is dramatic.
At a regional level, we can see population and urban transformation. But I’d like our map to do more. As I mentioned, I believe place-stories are important for how we think about history. And while this region-wide trends can be discerned, I also want to tell specific stories. I could envision, for example, a design like the one below.
The map here is an essay, tied to the visualization. The essay mentions specific places that can be interacted with: clicking a place zooms you to that spot and presents you with specific stories about human experiences. Scrolling through the essay likewise updates the map as you read, presenting readers with the spatial context of the history being argued.
What I want, in other words, are deeper maps: a way to tie together narrative and analysis with interactive maps. In this way, we can guide readers through spatial stories. How can we, with digital maps, tell more complete stories about urban renewal? About residential segregation? About infrastructure development? Politics, race, class, gender, culture, environment — interactive maps can help us tell these complicated stories, and perhaps even do so more easily than written text.
To present a useful and truthful picture, an accurate map must tell white lies. —Mark Monmonier, How to Lie with Maps
That isn’t to say that maps are perfect. There are lies, there are damn lies, there are statistics, and then there are maps. Maps, especially data maps, omit as much as they add. We must be cautious with the maps we maker. As Laurie Maffly-Kipp warnsa, maps are “still limited by data that is partial, ambiguous, and clearly slanted towards things that can be counted and people that traditionally have been seen as significant.” Our maps, in other words, too often end up being about things we already know a great deal about.
That isn’t to say that digital history is the perfect medium, either: digital history has plenty of hubris, and I don’t wish to amplify that here. But it does offer opportunities for experimentation. To take advantage of capacious historical sources, we need digital maps: we can work at new scales, display change over time, integrate maps with sources, craft new narratives. The maps we looked at today don’t solve all these problems, but it does permit us to start addressing the theoretical concerns. In effect, we’re able to make scores of maps, where before we could only make a few.
But we can start to build more detailed maps of Midwestern history, in ways that let’s us ask complex questions about space and identities. I hope you’ll agree with me that it’s time to create an atlas of the Midwest, and if we’re going to do this it needs to be digital.