"From manufacturing belt, to rust belt, to college country"

A Visual Narrative of the US Urban Growth

[Dani Arribas-Bel & Michiel Gerritse]

Results and code

What has shaped the US urban landscape? Probably, different forces worked at different magnitudes, times and locations. In this paper, we develop a methodology to disaggregate some of the engines of US city growth over time and across space. To understand the results, we propose a visualization approach based on what we term storyboards that creates an intuitive and dynamic narrative on the effect of several factors of urban success. This allows us to show that the role of growth engines greatly differs: the rise and decline of manufacturing was very localized; industrial specialization is counterproductive, particularly so in the 1990s; service sectors used to be a consumption amenity, but now serve as a production amenity; and highly educated cities unambiguously and increasingly attract firms in any part of the US. We also note that the arguments for visualization and lessons learnt in this exercise have greater implications for its role in the Social Sciences beyond this particular example.

This website contains the code and data required to reproduce the figures in the paper, along with some light explanation of the main insights and techniques that came out of this project. Data Preparation contains a description of the data used and STATA code to process it into the shape and form we need it for the visualization, including how to run the regressions for the Roback model on which we rely. Visualizations contains Python code to reproduce the figures presented in the paper. More importantly, the code is flexible and general enough that allows any user with basic Python skills to adapt it to different datasets or new questions.



This website is hosted in an open repository that contains both the website itself as well as the code and data required to reproduce figures and results in the paper. You can access its contents in the following two options:

Code and data can be downloaded separately from he Github repository. Either via the standard git workflow (clone or fork) or as a compressed file (link).


Creative Commons License
Results and code for "From manufacturing belt, to rust belt, to college country - A Visual Narrative of the US Urban Growth " by Daniel Arribas-Bel and Michiel Gerritse is licensed under a Creative Commons Attribution 4.0 International License.