Can this app predict vacancy? The city says it’s the ‘future of inspection’

The City of Philadelphia just built a new app to identify vacant properties or those that could be slipping into abandonment.

It’s an effort to get a better picture of vacancy across the city, one of the recommendations made by the task force the Mayor assembled after 2013’s Market Street building collapse. Here’s a snapshot of the problem: There are an estimated 25,000 vacant properties, according to 2010 Econsult report, and 40,000 vacant lots in Philly. For a rough comparison: there are 50,000 abandoned properties of America’s poster child of urban decay, Detroit.
The new app uses reams of city data from several Philadelphia agencies — property violations, water shutoffs, LIDAR imaging that can show things like a hole in a roof that an inspector might miss from the street level — to calculate the probability that a property or lot is vacant. The city’s roughly 50 property maintenance inspectors will use data from the vacancy app to prioritize their inspections.
“This is the future of inspection in Philly,” said Rebecca Swanson, the Department of Licenses & Inspections’ director of policy and legislative affairs.
A project between L&I and the city’s Office of Innovation and Technology, the tool has its roots in L&I’s vacant strategy unit, where then-L&I employee Stacey Mosley worked to identify vacant property indicators, Swanson said. That work was guided by a lot of hands-on research. This tech-enabled solution is helping L&I “take a leap forward,” Swanson said. (It reminds us of the predictive, data-driven work that the Philadelphia Police Department is doing.)

She stressed that the tool is still a work in progress. Inspectors will head out into the field to confirm or deny the tool’s accuracy and make it smarter.
It’s also not a replacement for the boots-on-the-ground efforts of city inspectors. Instead, the app aims to help the city be more proactive. Could it get the city to inspect a vacant property before it becomes unsafe?
The tool is especially helpful for a a concept as finicky as vacancy, Swanson said.
“Vacancy is such a snapshot,” she said. “A property can be vacant today and occupied tomorrow.”
L&I will also use the vacancy app in conjunction with another new app: one that helps identify properties that are likely to collapse or become unsafe. That model looks at things like whether the properties around a rowhouse have fallen down or if a property is close to a bus stop or a school and is more likely to be a safety hazard if it collapses.
Several departments across the city pitched in to invest more than $800,000 in the project, said Chief Geographic Information Officer Mark Wheeler. That cost covered taking LIDAR and aerial images of the city, plus analysis of that data. In its report, the building collapse task force said it strongly supported a LIDAR survey of the city and Wheeler said that was helpful in getting funding for the project.
(One dataset that isn’t included in the app? Property tax revenue data, a dataset that has appeared to be under lock and key, at least in a machine-readable format, and perhaps not just to the public. “We’re evaluating how we can use lien information if it’s available to us,” Wheeler said.)

Wheeler and GIS analyst Tim Haynes, who helped build the tool, showed off the vacancy app at this week’s well-attended GIS open house, held at the city’s Innovation Lab on the 16th floor of the Municipal Services Building. Roughly 50 people came to the event Tuesday evening to see what city agencies like the Water Department, L&I and OIT have been working on. Earlier that day, OIT held a similar event but for city employees, where L&I inspectors got to see the new tool.
Would the vacancy model ever be open to the public?
It’s still early in the process so that decision hasn’t been made yet, Wheeler said, adding that there are some downsides to releasing a predictive model.
“We’re not saying these are all the vacant properties or saying that the city is going to act on every one of these properties,” he said.