Rich Neighbors – Poor Neighbors: How Close Are They?

So the story behind this one is that I was working on a project that looked at the cost of housing and the median household income by county in California, and looking at the data I had made me curious about some other questions.  One of those questions was, ‘Are wealthy areas in the US generally close or far from poor areas?’

So first I had to get a little more concrete with my question so that I could apply some data to it (data generally dislikes vague questions), and after a little work, I came up with this.  ‘How does the distance between counties with high levels of wealth and low levels of wealth compare to the average distance between counties in any given US state?’  I landed here for a couple reasons.

  1. I couldn’t just compare the distance between HW (high wealth) and LW (low wealth) counties across states, because states are different.  If the average distance between HW and LW in Texas is greater than in Tennessee, we don’t really know if that’s just because Tennessee is smaller, so all the counties are closer together.
  2. I needed something to compare in order to determine if the distance between HW and LW meant anything, so I compared to average distance between counties.  This accounted for the difference in state size, and gave me a measure with which I could compare states.

So ultimately, the measure I used is a delta – the difference between the average distance between counties and the distance between HW and LW counties in a given state.

I also had to come up with a way of determining which counties to use for my HWs and LWs.  For simplicity sake, I use US Census Bureau data from the 2014 American Community Survey.  For HW I used the five counties in each state with the highest percentage of total households in that county in which total household income is greater than $200,000.  For example, if 10% of a county’s households had income over $200k, it would be ranked higher than a county in which 8% of the households had income over $200k.  Then I did the same thing at the other end of the scale for LW – I used the five counties in each state with the lowest percentage of total households in that county in which total household income is greater than $200,000.  So HW represents counties with a high concentration of high-income households and LW represents counties with a low concentration of high-income households.

“But wait!” you say,  “Having a low concentration of high-income households does not necessarily indicate that a county is poor.”  True enough.  So I did the same kind of calculation to identify what I called High Poverty (HP) and Low Poverty (LP) counties.  For HP I used the five counties in each state with the highest percentage of total households in that county in which total household income is less than $10,000.  For example, if 10% of a county’s households had income under $10k, it would be ranked higher than a county in which 8% of the households had income under $10k.  For LW I used the five counties in each state with the lowest percentage of total households in that county in which total household income is less than $10,000.  So HP represents counties with a high concentration of low-income households and LP represents counties with a low concentration of low-income households.

 

And what did I find?  Well, there were no bombshell revelations, but there were some interesting hints about how wealth and poverty are distributed in the states, at least relative to each other.  Here’s a summary view:

 

It turns out that overall, the distance between high-wealth and low-wealth counties is about 18 miles greater than you would expect, based on the average distance between counties across all states. In terms of high and low-poverty counties, the difference is even less at 5.5 miles greater than average.

Of course, there are all kinds of ways to slice and dice the data that are more interesting than the summary overview, so I’ve produced an interactive visualization that you can play with on my Tableau Public page.  To whet your appetite, here are a couple interesting tidbits that you’ll find on the more robust viz.

The part of the country you are in seems to make a difference.  The HW to LW delta is much greater in the eastern timezone than in the rest of the country.  However, that does not hold true for the HP to LP delta which seems much smaller difference across timezones.

The HW to LW delta is much greater in blue states than red states, while the HP to LP delta is greater in red states (though by a smaller margin).

HW to LW deltas appear to be strongly correlated with household size, and with total state population, though the same does not seem to hold true for the HP to LP delta.

 

Drop a comment below if you found anything interesting, if you have a question, or if you question any of my analysis.  I’m always open to the conversation and love to hear other’s perspectives.

 


Methodology Notes:

The county longitude and latitude data is from the 2014 U.S. Gazetteer Files from the US Census Bureau.  As of February, 2015, the file could be found on this page under the heading “counties”.

The household income data is from the 2014 American Community Survey 2014 from the US Census Bureau.  As of February, 2015, the data could be sourced from this page.

High/Low Wealth calculations: Counties sorted by % of total households in that county in which total household income is greater than $200,000.  The five counties with the highest percent of total households with total household income greater than $200,000 are labeled High Wealth (HW) counties.  The five counties with the lowest percent of total households with total household income greater than $200,000 are labeled Low Wealth (HW) counties.

High/Low Poverty calculations:  Counties sorted by % of total households in that county in which total household income is less than $10,000.  The five counties with the highest percent of total households with total household income less than $10,000 are labeled High Poverty (HP) counties.  The five counties with the lowest percent of total households with total household income less than $10,000 are labeled Low Poverty (LP) counties.

Connecticut, Delaware, Hawaii and Rhode Island were excluded from these analyses because the did not have enough counties to yield valid results.

Alaska is also excluded from the timezone analyses since it is the only state in its timezone.

 

Leave A Reply

Your email address will not be published. Required fields are marked *

Follow our Les Paul build thread and let us know what you think.
Holler Box