Philadelphia Housing Characteristics

Joy Payton
January 13, 2017

Deadbeat Investors

Deadbeat investor landlords (defined as owners who have 3 or more properties in property tax delinquency) are a real problem in Philadelphia. We've gathered data to help understand this problem from two sources:

  • Philadelphia's Open Data repository (tax delinquency)
  • US Census Bureau (poverty, income, and housing data)

We've organized this data into statistics by census tract, to understand the characteristics of neighborhoods with high levels of deadbeat investor delinquency. We used just one tax year, 2015.

Tax Delinquency Impact

Let's look at tax delinquency for all Philadelphia census tracts:

sum(PhillyHousing$InvestorDelinquentTaxTotal)

[1] 3533812

summary(PhillyHousing$InvestorDelinquentTaxTotal)

    Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
   16.21  3059.00  9865.00 13490.00 19090.00 87530.00

That's over $3.5m delinquency for 2015 alone! But some census tracts have very little delinquency, while others have quite a lot.

What correlates?

With the data at hand (housing characteristics, poverty and income, and tax delinquency), can we detect any relationships?

plot of chunk unnamed-chunk-3

Lots of Interactions

We can see by the correlation matrix that many explanatory variables influence each other, as well as predicting tax delinquency.

For that reason, it's a good idea for us to create a quick visualization Shiny web app that will allow us to plot any of those variables against each other.

For now, we're concentrating on discovering single-predictor models, so we'd like to generate plots like the one on the next slide:

Plots

plot of chunk unnamed-chunk-4

Check it out!