The same investors who once abandoned Philadelphia are now clamoring to get back in – and although new investment marks a reversal of fortunes for the City, it appears only a handful of neighborhoods have attracted this newfound attention.
Neighborhood change in Philadelphia has been a relatively slow process compared to New York and San Francisco, and this pace offers community stakeholders a unique opportunity to plan for equity in the midst of growth.
Click the image for a higher resolution visualization.
The goal is to develop an understanding of how neighborhood economies respond to new housing demand, and then use this intelligence to design more effective planning interventions.
These dynamics are part social science and part data science. With an emphasis towards the latter, this article visualizes the neighborhood change process in Philadelphia through several spatial and temporal datasets - many of which are open and free for download.
The term ‘gentrification’ is deliberately avoided here partly because of its pejorative connotation and partly because some of gentrification’s most interesting dynamics cannot be reduced to simple data visualizations.
The main indicator for this article is single-family home sale prices, which are well suited for describing neighborhood change. Real estate prices reflect the willingness to pay for different neighborhood characteristics, and if a shift in these characteristics typifies the neighborhood change process, then we should expect a shift in prices as well.
The image below visualizes home price trends in several Philadelphia neighborhoods between 2000 and 2013.
Click the image for a higher resolution visualization.
In some places, neighborhood prices follow the citywide trend but in others they deviate considerably. Why is this? What is it about certain neighborhoods that make them predisposed to new investment?
Two of the more important forces that dictate the willingness to pay for neighborhood housing are the quality of amenities and public services. Let’s take each in turn.
The above title image visualizes real price changes between 1999 and 2013 by neighborhood. The greatest change in prices can be found in and around Philadelphia’s central business district also known as “Center City”.
Not surprisingly, Philadelphia’s downtown boasts the City’s highest density of cultural and recreational amenities. The area features 11,000 hotel rooms, more than 400 arts and cultural institutions, more than 450 restaurants and “almost 180,000 residents in a dense, architecturally-rich, and walkable 17th century street-grid”.
Development in Center City is booming. The Philadelphia Business Journal reports $4.5 billion dollars’ worth of building projects are either currently underway or in the pipeline.
With such a critical mass of amenities and amenity consumers, it is not surprising that Center City home prices have surged in recent years.
The visualization below shows mean home prices as a function of distance to Center City.
Click the image for a higher resolution visualization.
Note the sharp decline in single family sale prices at the edge of Center City, suggesting that residents are willing to pay top dollar to live in and around Center City but that this premium quickly deteriorates with distance.
Home prices rise once again in the neighborhoods along the City’s outer periphery – a trend that may reflect moderate demand for neighborhoods that are comparatively less dense.
It is interesting to note that Center City prices were largely unaffected by the 2008 housing downturn while prices in the outer neighborhoods fell dramatically from their 2008 peak.
Amenities are important determinants of neighborhood economic development but public service quality is equally critical. Of many services, public safety and school quality are two of the most influential.
The below visualization plots home prices in a given year as a function of street crimes by neighborhood. Not surprisingly, the relationship suggests that more crime is associated with lower home prices.
Click the image for a higher resolution visualization.
The “R-Squared” value reflects the statistical correlation between the two variables on the plot. The statistic runs from ‘1’, which indicates perfect correlation, to ‘0’, indicating no correlation. Note that in the case of crime, the strength of the correlation remains steady over time. In comparison, let’s look at the relationship between home prices and school quality.
There is simply not enough space here to describe the multifaceted catastrophe that has befallen the Philadelphia public school system. We tend to think about schools as drivers of human capital development, but in Philadelphia, where only a handful of “regional choice” schools remain, good schools have become a driver of economic development as well.
Using newly released open test score data from the Philadelphia School District, the figure below visualizes the relationship between 3rd grade standardized reading scores and home prices by school attendance catchment zones over time.
Click the image for a higher resolution visualization.
Unsurprisingly, there is a significant price premium associated with good schools. The more interesting takeaway however, is that the R-Squared values suggest that this premium is becoming increasingly pronounced over time as more schools close and fewer good ones remain.
Thus far, the visualizations have illustrated the neighborhood change process at the neighborhood level (a seemingly sensible thing to do, no doubt). Perhaps more interesting however, is how this process unfolds at very microeconomic scales – even as small as an individual parcel.
Imagine that the spatial pattern of reinvestment resembles a wave with higher income households and higher home sale prices propagating across the urban landscape. If new demand is steady, this wave should continue unencumbered until it collides with a wall powerful enough to stem the tide.
These walls exist in cities. Some are natural – like Manhattan’s East River; some are erected to corral demand in one place; and some exist simply as artifacts of home buyer preference.
In 2001, the University of Pennsylvania opened the Penn Alexander School – a public school for which the University provides an additional per-student subsidy on top of the School District’s existing expenditure.
When it opened, an attendance catchment boundary was drawn around the school – a planned wall that splits the neighborhood into households who can attend the school and households who cannot.
The visual below plots inflation adjusted home sale prices as a function of their distance to the catchment boundary by year. Notice in the first two years before the school opened (the yellow line), prices on either side of the boundary were about the same.
Soon after however, as demand for the new school set in, sale prices on either side of the boundary began to diverge. The Penn Alexander case symbolizes how the introduction of a quality school where one previously did not exist can be a tremendous driver of neighborhood change.
Click the image for a higher resolution visualization.
As a second example, an eight block stretch of western Girard Avenue (east of the Schuylkill river) is a commercial corridor that has persistently prevented the wave of higher home prices from moving northward beyond Center City.
The visualization below plots home sale prices as a function of their distance to Girard Avenue. The Girard Avenue wall emerges over time and by 2004-2005, prices on either side are wildly divergent.
Click the image for a higher resolution visualization.
By contrast, let’s look at a wall in Philadelphia that appears to be crumbling. Low density commercial and industrial land uses along Washington Avenue on the west side of Broad Street have long separated the southern section of Center City from the Point Breeze neighborhood.
The visualization below shows that particularly in the mid-2000s, Washington Avenue acted as a wall preventing higher home prices from moving south away from Center City as the downtown boom fueled neighborhood change along its periphery.
The influence of the wall began to wane in the latter half of the decade as home buyers began to take advantage of inexpensive housing south of the Avenue but in close proximity to Center City amenities. While price differences still remain 2012-2013, the trend suggests that these differences will continue to wither with time.
Click the image for a higher resolution visualization.
Below is another view of the Washington Avenue process – this time in 3-dimensions.
Click the image for a higher resolution visualization.
This wave of reinvestment and its momentum and direction are influenced by small spatial variations in supply and demand conditions over time. Developing statistical tools to track this process requires a bit of creativity and a great deal of algorithmic insight.
No doubt, these metrics can help guide for-profit investors, but city governments and community developers could also benefit from their use, particularly when developing new interventions to help counter the potential threat of residential displacement.
These analyses can be used to forecast future neighborhood demand and help community stakeholders better allocate their limited affordable housing resources to ensure that when the wave hits, the benefits will be experienced by all.
In no way however, should these strategies preclude the need for an active dialogue between local politicians and their constituents. Instead they should complement the already burgeoning movement toward evidenced-based planning – one that we hope will help make cities more productive and equitable places to live and work.
Ken Steif is a Doctoral Candidate in the City & Regional Planning Program at the University of Pennsylvania where he studies housing policy and neighborhood change. He also consults and teaches courses in GIS and spatial/statistical analysis. You can follow him on Twitter @KenSteif.
Will Philly be a Model for the Future of Equitable Housing? »