I had the honor of giving the E. Willard and Ruby S. Miller Endowed Lecture in the Department of Geography at Pennsylvania State University on March 15, 2024. My talk was titled “Mapping Columbus’ Ghost Neighborhoods: Using AI and GIS to Create 3D Models of Neighborhoods Damaged by Urban Highways and Urban Renewal in the 20th Century.”
They even baked a cake!
In CURA’s Ghost Neighborhoods of Columbus project, we are applying machine learning methods to extract information from historical Sanborn Fire Insurance maps to generate realistic 3D models of how neighborhoods looked in the past. Machine learning – computer techniques that can find general patterns in data – can hep unlock the incredibly rich information in the building-level maps created for fire insurance underwriting purposes. Sanborn maps date back from the late 1800s until the 1960s and are available for over 10,000 cities and towns in the United States.
In our project, we are focusing on historically Black neighborhoods in Columbus that have been altered and, in many cases, damaged by deliberate actions such as urban highway construction, urban renewal and redlining practices. Our intent is to raise awareness and empower restorative justice by creating evocative experiences that connect with people in a visceral manner. A longer-term objective is to build a scientific database to which we can apply high resolution morphological analytics to understand the impacts of built environments – and their large–scale alterations – on social, health and environmental justice outcomes in these communities.
Here are some updates about our three main study sites to date: i) Hanford Village in 1961; ii) Poindexter Village in 1940; iii) Mt. Vernon Ave in 1951.
(All images can be clicked for larger versions.)
Our first study site is Hanford Village, a historically Black neighborhood in Columbus that was bisected and harmed by the construction of highway I-70 starting in the late 1960s. The image below shows our initial 3D model. The buildings are from 1961; the completed I-70 highway is superimposed on the image. Buildings in red were demolished for the highway and adjacent Alum Creek Drive (on the right side of the image, near the highway curve). Our results show that a total of 380 buildings have been demolished in these areas, including 286 dwellings, 86 garages, 5 apartments, and 3 stores.
Note that the buildings textures (exteriors) are plausible but not accurate. We applied simple rules within ESRI City Engine to generate these textures. In our current phase, we are studying historical and current day photos to build more accurate building textures. The image below shows a Google Street View image of a still existing 1940s Cape Cod home in the George Washington Carver Addition in Hanford Village with a 3D model of the building. We are currently working on making all of the buildings in our Hanford Village model this realistic and accurate.
Poindexter Village was one the first public housing project in Ohio and one of the first in the United States: Franklin Delano Roosevelt attended its dedication in 1940. Most of the buildings have since been demolished. The images below shows our 3D model of Poindexter Village, overlaid on a present day aerial image. The top image shows an overview; the bottom image shows building details. The only buildings that remain are the church (the building with the columns) and the two residential buildings to its left in both images.
A museum and visitor center is planned for the two remaining buildings. We are in the final stages of building this model. We are also working in partnership with the Ohio History Connection to explore ways to incorporate this model into the planned museum and visitor center, delivered via the web, onsite digital displays or a 3D printed tabletop model.
Mt. Vernon Ave in 1951 is our newest study site and time, supported by a seed grant from the Battelle Engineering, Technology and Human Affairs (BETHA) Endowment fund at The Ohio State University and in partnership with the Columbus Landmarks Foundation and the City of Columbus Near East Area Commission (NEAC). CLF and NEAC are pursuing historic district designation for Mt. Vernon Ave. Mt. Vernon Ave was the commercial heart of the Black community in Columbus in the mid 20th century; the construction of the I-71 highway severed this corridor from Columbus downtown. The image below shows a map of Mt. Vernon Ave in the Bronzeville neighborhood of Columbus. The dashed line indicates where the street no longer exists (the present day Columbus State Community College campus.)
Our initial study area is a three-block stretch of Mt. Vernon Ave between present-day Monroe St (to the west) and 20th Street (to the east). The images below shows a present-day aerial image with our study area demarcated at two different map scales. Based on our research, this was the largest concentration of commercial activity in 1951. It was also the site of an ill-considered urban renewal project in 1971 under the Model Cities program
The next three images illustrate the information extracted from the Sanborn maps. The first image shows building footprints with uses, overlaid on a present-day map. D = dwelling; F = flat (apartment); S = store (commercial); A= automobile (garage).
This image shows the number of stories derived from the Sanborn maps:
Finally, the building construction material from the Sanborn maps:
The animation below compares the present-day buildings in 3D to the 1951 buildings in 3D. Even with this basic, simple building models, we can see the sharp decline in density and loss of commercial activity after the urban renewal project.
We are currently conducting archival research to determinate accurate building textures, especially for the commercial buildings, and adding this information to the models. We are also working with Dr. Matt Lewis to develop an in-situ augmented reality experience of Mt. Vernon Ave in 1951.
A professional MC from EventRap hired as a discussant to “rap-up” our project presentation at the OSU TDAI/SI Interdisciplinary Research Fall Forum, November 9 2023
New paper: Lin Y, Li J, Porr A, Logan G, Xiao N, Miller HJ (2023) “Creating building-level, three-dimensional digital models of historic urban neighborhoods from Sanborn Fire Insurance maps using machine learning.” PLoS ONE 18(6): e0286340. https://doi.org/10.1371/journal.pone.0286340.
Abstract. Sanborn Fire Insurance maps contain a wealth of building-level information about U.S. cities dating back to the late 19th century. They are a valuable resource for studying changes in urban environments, such as the legacy of urban highway construction and urban renewal in the 20th century. However, it is a challenge to automatically extract the building-level information effectively and efficiently from Sanborn maps because of the large number of map entities and the lack of appropriate computational methods to detect these entities. This paper contributes to a scalable workflow that utilizes machine learning to identify building footprints and associated properties on Sanborn maps. This information can be effectively applied to create 3D visualization of historic urban neighborhoods and inform urban changes. We demonstrate our methods using Sanborn maps for two neighborhoods in Columbus, Ohio, USA that were bisected by highway construction in the 1960s. Quantitative and visual analysis of the results suggest high accuracy of the extracted building-level information, with an F-1 score of 0.9 for building footprints and construction materials, and over 0.7 for building utilizations and numbers of stories. We also illustrate how to visualize pre-highway neighborhoods.
Media
