Mapping History

This week, I used the Allmaps editor to georeference a 1855 map of Northfield, which shows land parcels archived at the Northfield Public Library. I uploaded the IIIF resource to Allmaps, drew a mask to trim unwanted edges, and then marked control points by matching features on the historical image to the contemporary map. With those correspondences, Allmaps computed the transformation and produced georeferenced tiles.

The most interesting part of this project was identifying reliable control points. Landmarks such as Central Park and the street grid were the easiest anchors. I even spent some time finding the correspondence for the historical locations of a few of my favorite riverside restaurants. It feels amazing to see how decisions made long ago still shape the present city space.

Reflection

This project also changed my view of digital humanities. I had assumed the main benefit of leveraging digital tools in the humanities was speed and expediency, using technology to streamline labor-intensive steps. From this project, I realize that accurately aligning historical imagery to modern maps can be meticulous work. When researchers georeferencing places with major topographic or urban change, establishing trustworthy correspondences demands even more effort.

On the map page, I noticed we can access the generated map as XYZ map tiles in addition to the raw image. We can also export links for the Georeference Annotation and the GeoJSON. The former lets others reopen and edit the georeferencing in Allmaps, and the latter provides the footprint that can be displayed in other maps.

Once we have georectified maps, I can envision many possibilities with the rectified map. By overlaying the historical map on a modern basemap, we can trace patterns of urban change over time. We can also examine shifts in topography, or broader landscape features, during the period. And by digitizing features from the map, such as parcels, streets, and landmarks, we can perform more detailed spatial analyses using the newly extracted data.

However, I think there are potential drawbacks to this method. To make effective use of georeferenced images, researchers must invest time and care in creating accurate labels. When historical maps are degraded or diverge substantially from the modern landscape, digitization and georeferencing become difficult. Because these sources are primarily visual, they offer limited support for causal inference without linking to additional, non-spatial data and methods.