The Wolfram|Alpha Blog is now part of the Wolfram Blog. Join us there for the latest on Wolfram|Alpha and other Wolfram offerings »
C. Alan Joyce

Studying Education and Income across the United States

October 10, 2012 —
Comments Off

First, my apologies: I didn’t quite follow through on my promise of a regular series of blog posts about American Community Survey data in Wolfram|Alpha. But when you’re trying to ingest all the world’s systematic knowledge… well, there’s a lot of competition for the top spot on your to-do list. So to make up for lost time, I’ll cover the remaining clusters of ACS data that you can currently explore in Wolfram|Alpha: education and income.

We’ve already got quite a bit of information about the United States’s educational system in Wolfram|Alpha, including student demographics for public and private schools and detailed financial information for school districts. With five-year estimate data from the American Community Survey (ACS), we’re now able to answer detailed questions about school enrollment and educational attainment across multiple levels of geography—states, metropolitan areas, counties, congressional districts, school districts, and individual cities and towns, as well as the nation overall.

So you might start with a simple question, like “What fraction of the US population is enrolled in college?” Based on 2006–10 ACS estimates, the answer is about 7.4% of the population age 3 and older, or nearly 22 million people. Clicking “Show details” and “Show chart” in the school enrollment pod will reveal a more complete breakdown of overall school enrollment:

What fraction of the US population is enrolled in college?

(Note, however, that the congressional districts we currently display are for the current [112th] Congress; we’ll be adding district boundaries for the 113th Congress soon, but ACS estimates won’t catch up to those boundaries for another couple of years. So unfortunately we’re in an in-between period where the data can’t tell us much about the current electorate.)

Or you might be interested in the relative size of specific school populations across many different regions—perhaps you’re curious about kindergarten enrollment by state or which California congressional district has the biggest fraction of people enrolled in grad school, or you’d like to compare the fraction of people enrolled in college in New York City versus Boston:

Compare fraction of people enrolled in college in New York City versus Boston

If you’re more interested in educational attainment—meaning the highest degree or level of school a person has completed—you can even ask questions about the population broken down by sex. Which county in Massachusetts has the biggest fraction of women with a PhD? How many male high school dropouts are there in Detroit? Or you could come up with a comparison of congressional districts that occupy two ends of the spectrum—one with the highest fraction of people with no formal schooling (California’s 20th), the other with the highest fraction of PhDs (Maryland’s 8th):

Educational attainment distribution in CA-20 versus MD-8

Fascinating stuff, but it’s the ACS data on income that I suspect will capture most people’s interest—especially with so much media attention over the past year on “the 99%” and questions of what constitutes a “middle class” income. So let’s start with the basics: according to five-year ACS estimates for 2006–2010, median household income in the US is just under $52,000. Try that basic input in Wolfram|Alpha, and you’ll see the distribution of household incomes in tabular form, as well as in a helpful visualization.

Median household income in the US

Keep in mind that the bins at the lower end of the scale mostly cover ranges of $5–$10,000, while the upper bins cover ranges of $25–$50,000. It might look to some people like the main concentration of households is right around the $100,000 mark, but a quick question about the fraction of households earning below $99k should set them straight: nearly 80% of households in the US fall below that line.

So where are most of the top 20% of households concentrated? Ask Wolfram|Alpha to show you US states by percent of households earning $200k, and you learn that number one is… New Jersey. I found that result a bit surprising, so I took a moment to compare New Jersey with the United States overall on this count, and sure enough, the density of top-earning households in New Jersey is nearly twice that of the nation at large:

Compare percent of households earning $200k in New Jersey and US

But then I went a step further and asked about counties in New Jersey over that threshold—and found out that Hunterdon County topped the list, with more than 17.5% of households bringing in more than $200k per year, and a median household income of more than $100,000.

But there’s still plenty of useful and interesting information to be gleaned from the data we’ve added so far. From a look at the fraction of adult women below the poverty line by state to the county with the highest median household income, there are any number of different ways to analyze income levels in the US. Or you can chain queries about ACS data together with other stats in our knowledge base: ask about the unemployment rate in states with the highest Hispanic population fraction, or the median home value in Arizona counties with the largest American Indian population. Or simply ask Wolfram|Alpha about any city, state, or other geographic area in the US—those results now include a summary view of key ACS-based statistics, presented alongside other key socioeconomic information like the cost of living, home prices, and unemployment (the data below is from Wolfram|Alpha results for Chicago):

Chicago, IL

You may have noticed that the Census Bureau just released ACS one-year estimates for 2011, but those are only available for larger geographic areas. When the detailed 2007–2011 estimates come out in a few months, we’ll have enough data points available for most geographic areas to start visualizing some short-term trends for most of these properties. So keep watching the blog for updates… and in the meantime, keep writing in with comments, questions, and suggestions for other detailed socioeconomic datasets you’d like to see in Wolfram|Alpha.

1 Comment

How would you find the top 1% of per capita income?

Posted by Dillon Korman October 10, 2012 at 3:53 pm