Elders’ stories are bringing digital models of lost communities to life

An error occurred while preparing your download

Elders' stories are bringing digital models of lost communities to life

At 5-year mark, Ghost Neighborhoods project aims to pick up speed

Some of the 3D digital models created by researchers to depict lost neighborhoods in Columbus, Ohio, tell a clear story by placing the "ghosts" of houses that were demolished for freeway construction atop the roadways now occupying that land.

But by talking to the people who lived in these communities, scientists are filling in historic gaps that technology can't fill, adding trees, cars, people, photographs and life stories to the digitized infrastructure that was an initial focus of the work.

The research team from The Ohio State University has published an update on the Ghost Neighborhoods of Columbus project in the International Journal of Digital Humanities, reporting on the technology workflow, including solutions to bottlenecks they've encountered, and acknowledging a slow start to community engagement that hums along nicely these days. The team aims to put both the digital work and story collection on a faster track.

Launched in early 2021, the project has digitally reconstructed three historically Black communities that were fragmented or demolished to make way for new freeways under the 1956 National Interstate and Defense Highways Act. Prompted by a Columbus Dispatch article describing the heartbreak of highways being built through neighborhoods, the researchers set out to recount the human impact of urban renewal decisions of the past, hoping that these historic insights could have a positive impact on any future development.

"Doing this in a community-engaged way is not trivial," said senior study author Harvey Miller, professor of geography and director of Ohio State's Center for Urban and Regional Analysis (CURA). "What we're trying to do is show how we went about making sure that these models not only have validity, but that they actually reflect what people remember, what these neighborhoods used to look like - and also add value to the community."

While the project started as a purely scientific pursuit, the digital models and accompanying stories have found a future permanent home. The Poindexter Village African American Museum, set to open in 2028, will have a room dedicated to the Ghost Neighborhoods, said Shelbi Toone, project director for the museum, which will be the newest site within the Ohio History Connection system.

Poindexter Village, an early U.S. public housing project and the only one dedicated to African American residents, is one area recaptured in 3D models by CURA researchers. Only two of the 35 original apartment buildings still stand, and those structures, connected by an atrium to be built between them, will house the new museum.

The other communities modeled in the project are Hanford Village as it looked in 1961 and Mount Vernon Avenue in 1951. Much of Hanford Village, established in the early 1900s and expanded during World War II to house Black veterans, was razed for construction of I-70 in the 1960s. Mount Vernon Avenue was a busy commercial corridor in the mid-20th century before its connection to Downtown was severed by construction of I-71.

Miller and colleagues reported in 2023 on using machine learning to extract data from Sanborn fire insurance maps to create digital models. Since then, the researchers have focused on efforts to automate the generation of 3D models that are historically accurate, which has been a taller order.

Graduate student Tshui Mum (Summer) Ha has been working on ways to streamline and accelerate the 3D modeling process. This includes finding similar groups of buildings based on their shapes and semantic attributes - such as the number of stories, building use and construction materials - identifying their architectural elements, and developing templates for 3D modeling.

"Many existing studies have been using just the exterior walls of the buildings for clustering. We realized that including semantic attributes of those building footprints is a huge improvement compared to just looking at the exterior walls. We also found that having interior lines of the buildings may hold additional value to further improve the clustering results, and this will be our next step," Ha said.

"The better we can cluster these buildings, the better we can understand their distinctive architectural styles and the more accurately we can generate plausible 3D models at scale. We are benchmarking these methods that, hopefully, we would be able to identify the most effective approach to help us speed up the modeling process."

The goal to speed things up is, in part, related to making community engagement as productive as possible.

To date, the team has been building historically accurate models and then asking for elder feedback in a variety of ways, often attending community meetings with stickers and storyboards that attendees use to add finishing touches that bring the models to life.

"For all three neighborhoods, we're adding more realistic backgrounds like brick streets and grass so it's not just a map," Miller said. "What we'd like to do is generate a first approximation of a neighborhood using accelerated techniques, engage people in the process early on, and come back later with a more polished product. Not too far in the future, we expect to be able to take a laptop with 3D models, ask people, 'What do you think about this building?' and make adjustments on the spot."

Toone, the museum project director, connected the CURA team with James Preston Poindexter Foundation historians and elders to record oral histories related to landmarks and neighborhoods in the models, which will also be featured in the museum.

"I thought this was something that people need to see and experience. It needs to be tangible to the public," said Toone, also a co-author of the paper. "There's a lot of narrative to tell when you pull back to 1940 and really talk about the thriving neighborhood that used to be there. And I think it gives people a true sense of preservation and understanding of what was here through the technology since so many of these buildings have turned over, been demolished or changed."

Soon, the team plans to ask elders for suggestions for other lost-city themes - ideas for churches and the funk music scene have been floated so far - that should be similarly digitally re-created with accompanying stories.

"One important thing is that we're trying to establish a platform for people in the community to tell their stories," Miller said. "They're not our stories - they're theirs."

This work was funded by the U.S. National Science Foundation and the Battelle Engineering, Technology and Human Affairs endowment at Ohio State.

Additional Ohio State co-authors include Ningchuan Xiao, Matthew Lewis, Mostahidul Alam, Oliver Gwynn, Michelle Hooper, Karyn Kerdolff, Gavin Levine, Yuantai Li, Mahnoush Mostafavisabet, Joshua Sadvari, Josie Stiver, Jordan Swaim-Fox, Shubh Thakkar, Ahmad Ilderim Tokey and Di Wang.

More Ohio State News