Identify potentially unsafe buildings with this tool

Local data scientists Ken Steif and Stacey Mosley want to help governments use data to make better decisions.

Steif has done it before with a project that helps cities identify bike share opportunities through his consultancy Urban Spatial Analysis.

The duo’s most recent project identifies buildings in Philadelphia that are potentially unsafe. It’s a tool that can be used by inspections agencies to better allocate their resources, according to Steif. In other words, it’s one way for building inspectors to get ahead.

Steif tapped Mosley for the project in the fall. Check it out below. Proof of concept

Mosley said she first saw the need for something like this when she was working on the Vacant Property Strategy for Licenses and Inspections. There, she said she saw compliance rates increase dramatically with the aid of data analysis.

“When a department oversees a city of our size, with almost 600,000 properties citywide, data analysis can efficiently support the identification of potential problems and the management of the consequent workload,” she said.

Here’s how the tool works: Steif said they started with the hypothesis that there’s a relationship between unsafe buildings and neighborhood conditions.

“We try to model this relationship statistically by using a slew of open data to predict actual building demolitions,” he said. “We then come up with a series of metrics that describe how well we’re predicting and generate, for each parcel citywide, the probability that it might be unsafe.”

The question is, Steif said, would prioritizing inspectors using a tool like this make them more efficient? Regardless, the next step is to work with a city interested in using the tool, improving on it and putting it to work.