On September 30th, Penn IUR hosted the first MUSA Brown Bag lunch talk of the year, highlighting GIS work happening around Philadelphia. Ken Steif, Director of Penn’s Masters of Urban Spatial Analytics (MUSA) program, kicked off the event by welcoming the attendees and introducing Grant Ervin, Deputy Geographic Information Officer for Enterprise Services at City of Philadelphia, and Tim Haynes, GIS Analyst for the City of Philadelphia. They presented their recent work launching the Philadelphia vacancy model, which uses reams of city data from several Philadelphia agencies to calculate the probability that a property or lot is vacant. Pulling from a variety of data sources across Licenses & Inspections (L&I), Office of Property Assessment (OPA), the Philadelphia City Planning Commission (PCPC), the Philadelphia Redevelopment Authority’s land management software (LAMA), and the local utility companies, they identified a variety of indicators that help determine the likelihood of vacancy. Their office specifically developed two models: one for indicating the likelihood of vacant land and one for predicting the likelihood of a vacant building.
Showing a map of a Philadelphia neighborhood, they demonstrated how the model allowed them to visualize all predicted vacant lots, which were represented on a color scale of high to low likelihoods of vacancy. A second map visualized the same information for vacant buildings. By associating certain indicators with a certain amount of points, Ervin described, they can then develop a point total that indicates the likelihood of vacancy for each parcel or property. Scores of just 50% or higher were typically verified to be vacant buildings or lots, since many departments have outdated or conflicting information. By testing their hypothesized indicators primarily through the use of aerial imaging, they were able to narrow down their list of indicators. Ervin and Haynes also introduced new tools they have begun exploring as enhancements to the model. Their team completed Cyclomedia custom street views for all of Philadelphia as part of a pilot project that provided the department with comprehensive and up-to-date street images of the city. This project provides more accurate data than is available from Google, which often features images from outdated or multiple timeframes. Haynes also noted that they will soon be receiving new building footprint data, which has not been updated since 2004, that will allow them to test the results of their vacancy predictions for any buildings that are no longer present. They have also explored the use of LIDAR (Light Detection and Ranging), which produces precise, three-dimensional maps by surveying distance from a satellite or aircraft to the earth or earth’s surface characteristics. The data difficult to work with if not buildings are not classified correctly, so they are evaluating continued use of the technology. Both argued that the model is helping identify potential vacancy for investigation and action. While they pointed out that inspections and other activities leveraging the model have been limited to-date, many departments have plans to employ the data to assist with setting priorities for inspections, demolitions, investments, and a wide variety of other city functions. Steif wrapped up the event by thanking the speakers and reminding the audience that the talk is part of a monthly series hosted by MUSA.