What does gentrification in Philadelphia look like?
“High-rise, modern apartment buildings.”
“(A) modern look that’s so out of place with our traditional row homes that have been here for a hundred years.”
“Six- to seven-floor high-rises with garages in the basement. They charge an extra $200 to park.”
“Gray, industrial looking.”
“The houses are ugly as heck. No architectural style. They’re probably two-bedroom, some probably one. And they usually put a deck up. It’s not geared for kids or families. A lot of steps.”
These are some of the descriptions that longtime residents of gentrifying neighborhoods in Philly used to describe the new construction popping up around them.
We are Ph.D. candidates in architectural engineering and geography, environment and urban studies at Drexel and Temple universities in Philadelphia. Working with a multidisciplinary team of professors and students, we recently developed a new way to map gentrification in Philly neighborhoods using a combination of accounts from longtime residents, Google Street View images and machine learning.
Signs of gentrification in Philly include new buildings that don’t fit the surrounding architecture.
Jeff Fusco/The Conversation U.S., CC BY-SA
Using AI to spot gentrification
Our team posited that the best source for knowing what gentrification looks like comes from the perceptions of longtime residents in gentrifying neighborhoods.
So we held focus groups in three rapidly gentrifying neighborhoods – one in Northeast Philadelphia and two in the River Wards section north of Center City and along the Delaware River.
We asked residents to identify the visual cues of building designs, materials, colors and landscaping choices that they associate with gentrification.
Many of these residents could recount, in great detail, the exact street intersections where they saw gentrification-related development occur over the decades.
We corroborated each location they identified through historical Google Street View imagery. By examining the exteriors of these buildings, we could expand upon the more generalized language used in the discussions, such as “modern” or “boxy,” to more architecturally specific language, such as “presence of bump-out windows” and “increased floor area ratio,” which is a measure of how much of the surface area of a land parcel a building takes up.
When pulling panoramas of residential building exteriors from Google Street View, we looked at two distinct time periods: 2009-13 and 2017-21.
AI is getting better at spotting the visual signs of gentrification. Researchers refer to AI systems that categorize scenery according to certain characteristics, like seeming “gentrified” or “not-gentrified,” as “deep mapping” models.
Deep mapping models use neural network algorithms, which can pick up on patterns in big datasets. The particular model we used is…
