From the 1850s through the 1960s, expositions were the world’s premier spectacles of technology, industry, and culture. Staged in cities around the globe, from Nashville to Osaka, and organized around themes relating to progress and imperialism, expositions attempted to contain and represent all the products and processes of modernity. They invited visitors to experience, in a single place, a complete image of an interconnected world and a globalizing economy—in Mircea Eliade’s words, an imago mundi.
Expositions were simultaneously events and places: massive mini-cities built quickly, enjoyed briefly, and destroyed promptly. At the end of their runs, Paul Greenhalgh notes, they “disappeared into an abrupt oblivion.” Temporariness was essential to the form and function of an expo. Most obviously, because exposition structures had to be built rapidly, they were invariably made of cheap, modular, and disposable materials. From the 1890s onward, “staff,” a paintable plaster that plausibly mimicked all manner of stone, reigned supreme. More importantly, however, expos intended to describe history rather than become a permanent part of it. By submitting to planned self-destruction, expos largely avoided the slow unfolding of ruination and indifference that would undermine their potent yet moment-specific visions of the modern world.
Preserved as memory and image, in carefully crafted paraphernalia and souvenirs rather than in crumbling facades, they could be immunized against political or economic crisis, war, or abandonment by future generations. Their amnesty from physical decay ensured—or, at least, was intended to ensure—that the cultural values they embodied would never become visualizable as decrepit or passé.
Despite their ephemerality, expos were spaces of intense human activity—building, dwelling, gazing, visiting, experiencing, embodied practice and performance, the construction of bourgeois subjectivities—that imbued their sites with cultural power and meaning. From remnant art palaces left standing at fair’s end, to subtle road curvatures and parkland geometries, expositions remain encoded in the landscapes where they once put down shallow roots. Some, for example, spurred massive and lasting architectural and infrastructural development in their host cities, which can still be seen quite readily. The Seattle Space Needle and the Nashville Parthenon are among the more monumental examples of expo structures built to last. Other expo sites were redeveloped into indistinguishable forms or abandoned forever. The northern half of the 1901 Pan-American Exposition site, in Buffalo, is now a residential subdivision. Chicago’s Jackson Park, which was dramatically terraformed and built up for the 1893 World’s Columbian Exposition, has been for the better part of a century a scarcely-visited woodland and swamp—still bounded, nonetheless, by the artificial perimeter created by its architect, Frederick Law Olmsted.
Yet even if the scarcity of ruins on former expo sites prevents us from enjoying the sort of multisensory interaction that is the hallmark of modern-day heritage tourism, traces of expos can still be seen—and explored—through a digital humanities approach.
Using geospatial information systems (GIS), I illustrate in “Mapping the Ephemeral Vistas” how expos’ forms and structures continue to be perceivable in modern urban landscapes.
In this project, still a work in progress, I georeference historic exposition maps to modern maps, allowing us to observe how, where, and why these transitory landscapes endure. Because their official ground plans and souvenir maps often contain a great deal of contextual information—routes of railway lines, outlines of civic buildings, streets incorporated into fairgrounds or used as boundaries—overlaying maps from past and present reveals many of the cultural and infrastructural changes each city has undergone in the intervening years. There is of course an element of simple amusement here: discovering, for example, that standing on the former site of the beloved and monumental “Electric Tower” from the 1901 expo in Buffalo is now an unassuming block of split-level homes. But georeferencing these maps does more than enable these incidental, if marginally informative, discoveries. It enables us to consider, on a much larger scale, how our civic spaces are inveigled in a persistent cycle of invention, destruction, and reimagination; and how every cultural process, regardless of how fully it attempts to erase itself, leaves useful remnants that provide a window into the past.
I prototyped this work in Google Earth Pro, which allows users to superimpose so-called Image Overlays (PNG, TIFF, and JPEG) on its maps. After “clamping” the image to the ground, the user can manually morph an image file of a historic map—rotating it, stretching it, pulling it in various directions—until it visually corresponds with the cartographic features of the modern map. In terms of technical difficulty, this process is perhaps only only one or two steps above “drag and drop.” Yet it nevertheless requires enough familiarity with the modern cityscape that the often-subtle patterns and similitudes connecting past and present urban features—roads, parks, shorelines—can be discerned and matched.
Predictably, Google provides a rather engaging user experience, complete with cinematic “fly-through” animations and the ability to very quickly toggle informational layers such as modern roads, borders, and place labels. This video provides a sense of Google’s UX, within the context of this project.
Despite its ease of use and its suitability for quick prototyping and geographic research, Google Earth Pro is, in the broader scheme of things, a highly inappropriate tool for this project. Most significantly, its image overlays are sandboxed: there is no consistent and effective way to save and share them as files containing relevant GIS metadata that can be unpacked and used for other applications. In other words, manipulated image overlays, however visually compelling they may be, do not generate anything that could be reliably evaluated by anyone but their creator. Overall, despite its slick UI, Google Earth Pro fails the fundamental “open source” test by which I tend to evaluate others’ DH projects. It’s eye candy in a black box.
QGIS, the open source Mac variant of ArcGIS, is an undoubtedly superior tool for this project. In spite of its creaky graphics engine, which necessitates redownloading map tiles nearly every time the end users zooms and pans, QGIS allows for the creation of incredibly accurate—and, importantly, exportable and shareable—georeferenced files. My preferred file format is GTIFF—essentially, a high resolution image file that contains geographic (or cartographic) data embedded as tags. According to Mike Ruth,
GeoTIFF is a metadata format [that] provides geographic information to associate with the image data. But the TIFF file structure allows both the metadata and the image data to be encoded into the same file. GeoTIFF makes use of a public tag structure which is platform interoperable between any and all GeoTIFF-savvy readers. Any GIS, CAD, Image Processing, Desktop Mapping and any other types of systems using geographic images can read any GeoTIFF files created on any system to the GeoTIFF specification.”
As a base layer in QGIS, I rely on Mapbox‘s default, open source satellite view, which includes recent imagery and nearly every modern street label imaginable. I then employ the QGIS georeferencer plugin, first importing the JPEG, TIFF, or PNG file of the exposition ground plan; then determining useful ground control points (GCPs), particularly street intersections; and finally applying a predesigned transformation algorithm (thin plate spline, or TPS) that coordinates—and thus overlays—corresponding GCPs. This algorithm is known for being particularly effective at manipulating low-quality reference imagery, thus helping to overcome one of the larger obstacles in this project: the scarcity of good digitized maps for many of the more arcane expos. 1901 South Carolina Inter-State and West Indian Exposition, anyone?
This screenshot provides a sense of the QGIS georeferencing interface.
On this map of the 1905 Lewis and Clark Exposition in Portland, Oregon, red dots indicate the GCPs. The fact that nearly all the streets surrounding the southern perimeter of the fairgrounds still exist is a huge boon for the accuracy of this particular GTIFF. The “Residual (pixels)” column in the GCP table is particularly important: it represents, essentially, a confidence level in the maps’ cross-referentiality based on a comparison of the proximity of the GCPs on the historic map and the proximity of the GCPs on the modern reference map. The closer each GCP’s residual is to zero, the better. In this case, the very elegant hand-drawn map from 1905 is a quite excellent fit for the modern, Pseudo-Mercator (EPSG 3857) projection map provided by Mapbox. Very little perspectival manipulation is necessary to get the two to “click.”
Once the transformation algorithm is executed, the maps are effectively merged in the graphical interface. Underneath the hood, the map image file is now embedded with geospatial metadata, ready to be exported as a GTIFF (and subsequently imported as a raster tileset in ArcGIS, QGIS, Mapbox or another open source GIS interface).
Overlaying these maps instantly yields some fascinating points of historiographical departure. We see, for example, that Guild’s Lake, which formed the northern terminus of the Lewis and Clark Exposition, has since been replaced by landfill and transformed into a massive industrial park, its anonymous “spatial products”—warehouses, logistical hubs, etc.—staring blankly up at the satellite. We see, too, that this arc of the mighty Willamette River has been filled in as well, owing to corporate and industrial imperatives. Yet much of the surrounding landscape has remained the same. Upshur, Thurman, and Vaughan Streets are intact, as is much of the woodland to the southeast. In spite of one century of significant industrial development, during which the city of Portland’s population grew sixfold, this landscape has retained at least some of its original form and character.
Getting these maps to speak to each other opens up a world of inquiry, and reveals the continuities and ruptures—the opulent industrial palaces replaced by indistinct factories and warehouses—that are the hallmark of globalized, industrialized modernity.
Georeferencing illuminates the erasures and augmentations that shape our experience of space in modern life. It reminds us that the history of our built environment is not just what is observable and palpable but also what is hidden and forgotten. It asserts that even that which refuses to endure can often endure, in ways that demand our attention and comprehension. And, finally, it enables us to return to these landscapes in person, as tourists and archaeologists, and contemplate the self-effacing permanence of the ephemeral.
This .zip file, which I encourage all comers to download and play around with, contains fifteen GTIFF maps of American expositions. In future iterations of this project, global expositions will be georeferenced as well.