Ohio State University Researchers Use Oral Histories to Restore Digital Models of Demolished Black Neighborhoods

Researchers at Ohio State University merge 3D technology with elder stories to digitally restore Black communities destroyed by 1950s highway projects.

By: AXL Media

Published: Mar 23, 2026, 10:00 AM EDT

Source: Information for this report was sourced from Ohio State University

The Reconstruction of Lost Urban Landscapes

Researchers at The Ohio State University are expanding the scope of the Ghost Neighborhoods of Columbus project by utilizing 3D modeling to resurrect communities erased by the 1956 National Interstate and Defense Highways Act. While the initiative initially focused on the technical placement of residential structures atop modern freeway maps, it has evolved into a deeply human endeavor. According to Harvey Miller, a professor of geography at the university, the team is now moving beyond simple infrastructure to include specific environmental details like trees and period-accurate vehicles. This shift reflects an effort to ensure that the digital simulations serve as valid reflections of collective memory rather than just cold architectural data.

A Permanent Home for Digital Heritage

The digital recreations have secured a future outside of the laboratory through a partnership with the Poindexter Village African American Museum, which is scheduled to debut in 2028. This facility will occupy the remaining two buildings of a thirty-five building public housing complex that once served as a cornerstone for Black residents in Columbus. According to Shelbi Toone, the project director for the museum, the inclusion of these 3D models will provide the public with a tangible sense of what was lost during decades of urban renewal. The museum intends to use a dedicated room to display these "ghost" environments, allowing visitors to experience the thriving nature of the 1940s landscape before it was fragmented by industrial progress.

Technical Innovations in Historical Mapping

To manage the massive scale of these lost districts, the research team is implementing machine learning and advanced clustering techniques to automate the creation of accurate building models. Graduate student Tshui Mum Ha explained that the team is now looking beyond exterior walls to include semantic attributes such as building materials and internal floor plans to improve accuracy. By identifying recurring architectural styles from historic insurance maps, the researchers hope to develop templates that can generate plausible 3D models at a much faster rate. This technical efficiency is designed to allow the team to move from raw data to a recognizable neighborhood structure that can be presented to former residents for immediate verification.