In order to convert a gorgeous nineteenth century atlas into usable data, each exquisitely crafted sheet must be subjected to what might seem an utterly reductive - and possibly violent - process. The first stage of the process is warping. The second is disaggregation.
The georeferencing process involves creating links between specific locations on a basemap and related locations from an image file. Once a certain number of links are established, GIS software can run an algorithm that assigns spatial coordinates to the entire surface of the image.
In order to put the Piadyshev atlas into dialogue with modern basemaps such as Google Satellite or Open Street Map, we had to determine scale and projection. This was not a task for the faint of heart: it turned out that the atlas sheets were executed not only at different scales, but in different projections. This makes sense given the longitudinal breadth of the empire, but it made the work of stretching and fitting together the provinces into a kind of mosaic incredibly difficult. The MadMappers team shrinks from no challenge, however. They opted to follow Piadyshev's lead and work from an equidistant conic projection, customizing the central meridian to fit the Eurasian landmass. They created east and west mosaics and georeferenced them with extraordinary care and expertise. As a result, the derived layers offer the user a remarkably accurate view of the space of the Russian Empire as represented on the most sophisticated atlas of the early 19th century. Careful users will recognize that the historical representation of space does not correspond exactly with that produced using modern satellite imaging and cartographic science. The 'distortions' or 'inaccuracies' of the Piadyshev atlas are minimal given the scope and complexity of the georeferencing process; moreover, they present users with an opportunity to explore central issues in the history and science of map production.
Disaggregation (feature extraction)
In simplest terms, we set out to distill the geometries - the points, lines, and polygons - that define this complex cartographical object. We extracted each feature of the atlas, wrenching villages away from the roads around which they clustered, abstracting whole layers of information from their assigned symbologies, mercilessly lopping off titles and legends and any other evidence of the material context of each page.
No digital object can replace the materiality of the printed atlas. Gone are its heft and odor, the texture of each page, the sound of each page turning. The beauty of disaggregation is that it recreates the original, physical, map as a complex digital object: an object that can be studied and annotated, surely, but one that can also be manipulated, transformed, and analyzed using an entirely different set of tools.