Historical maps contain rich information about buildings, land use, and transportation networks, and provide novel insights into urban and spatial economics. However, this information is encoded into pixel values, and cannot be readily used for analysis. This paper investigates the evolution of French urban areas from a historical perspective. We implement a fully-convolutional neural network to extract pixel-level information from multiple collections of digitized historical maps covering mainland France in the 18th, 19th and 20th centuries. Using the extracted buildings, we define consistently urban areas and analyze their trajectories along the urban hierarchy. Our model uses an encoder-decoder structure and performs with remarkable accuracy and stability. The model successfully captures complex interaction patterns among neighbouring pixels values at various spatial scales and proves robust to considerable representation heterogeneity, both within and across maps collections, as well as severe class imbalance. This approach is efficient, scalable and readily transferable to other historical maps with minimal manual labelling. Our findings highlight increasing urbanization with fewer and larger urban areas, consistent with agglomerations economies. Disaggregate analysis reveals significant heterogeneity, with the persistence, emergence, and disappearance of urban areas.

Presenter: Clément Gorin, Postdoctoral Researcher at the University of Toronto 

Light refreshments will be available (in person)

This event will be held in person (Blackburn Room, 4036 Robarts Library) and virtually. Directions to the Blackburn room can be found at this link.  

Date: February 28th, 2023

Time: 2PM - 3PM