Wednesday, January 11, 2017

NY City Public Schools, and what they might tell us about the SAT

Recently, I received a message from Akil Bello who pointed out a data visualization he had seen.  It was originally posted to Reddit, but later was edited to eliminate the red-green barrier that people with color-blindness face.  The story was here, using a more suitable blue-red scheme.

There's nothing really wrong with visualizing test scores, of course.  I do it all the time.  But many of the comments on Reddit suggest that somehow the tests have real meaning, as a single variable devoid of any context.  I don't think that's a good way to analyze data.

So I went to the NY City Department of Education to see what I can find.  There is a lot of good stuff there, so I pulled some of it down and began taking a look at it.  Here's what I found.

On the first chart, I wanted to see if the SAT could be described as an outcome of other variables, so I put the average SAT score on the y-axis, and began with a simple measure: Eighth grade math and English scores on the x-axis. Hover over the regression line, and you'll see an r-squared of about .90.

Scientists would use the term "winner, winner, chicken dinner" when getting results like this.  It means, for all intents and purposes, that if you know a high school's mean 8th grade achievement scores, you can predict their SAT scores four years later with amazing accuracy.  And--here's the interesting thing--the equation holds for virtually every single school.  There are few outliers.

Ponder that.

But critics of the SAT also say that the scores are reflective of other things, too; an accumulation of social capital, for instance.  So use the control at the bottom to change the value on the x-axis.  Try economic need index, or percentage of students in temporary housing, or percentage of the student body that are White or Asian. The line may go up (positive correlation) or down (negative) but you'll always see the schools with the highest scores tend to have the characteristics you'd expect.

Jump to the second tab.  This is more a response to the Reddit post: The top map shows the ZIP codes and a bubble, indicating the number of schools in that ZIP.  The bottom map shows every school arrayed on two poverty scales: Economic Index and Percent in Temporary Housing.  The color shows the mean SAT score in the school (Critical Reading plus Math, on a 1600-point scale.)  Purple dots represent higher scores.

Use the ZIP highlighter, and you'll see the top map show only that bubble, and the bottom will show the schools in it.

Got the lesson?  Good.  Now, think about why the colleges with high median test scores a) have them, and b) tend to produce students with high GRE and MCAT and LSAT scores,  and c) point to excellent outcomes for their students.

And let me know what you think.






Wednesday, January 4, 2017

The Outlook in Illinois

Much of what I post here is slightly modified from what I use at work, and this is no exception.  Here at DePaul (like most universities) the biggest single slice of enrollment comes from our own state, and it's important to know what's going to be happening to the student markets in the future.

So I downloaded data from The Illinois State Board of Education showing enrollments for two years: 2010--2011 and 2015-2016 to see how things have changed over time, and to get a glimpse of the future.  This is a more granular look than the WICHE data I visualized recently, but it's also not actual projections going forward, but rather just numbers; projections require a lot of time and mathematics skills, neither of which I have.  I would have liked to gone deeper and farther with this, but the data are messy, and even things like School District IDs have changed over time.

There are four views using the tabs across the top: First by region, then county-by-county, and then a scattergram showing each county by both percent change and numeric change over time.  On each, make a choice at the of the page to change the data displayed: You can look at total pre-K through 12 enrollment, if you like (the default view) of you can change to show grade-level enrollments, or by ethnicity or low-income status.

Finally, the last tab shows individual schools.  You can type part of your school's name in the drop down box to start filtering, but be sure you find the county as well as the school.  If you're going to be looking up "Lincoln," you've got a lot of work to do!  Also, some schools have their name listed slightly differently in different years, and if your school is one of them, you won't get two years of data showing.

Please note: The data are not granular, so you can't combine variables (for instance, low-income students in 8th grade.)  And, I've excluded small numbers from the analysis (students in juvenile detention centers, or public school students being education at other sites.)

But it's still interesting, I think, especially if you drill down a bit using the filter at the top.

What do you see? Leave a note in the comments.




Friday, December 16, 2016

Medical, Law, and Dental Degrees, 1955-56 to 2013-14

You can look at a lot of places on this blog to find the story of women and the increases in educational attainment over time, but perhaps none is so compelling as this one.  It was very rare for women to have college degrees in the 1940's and 1950's, but even rarer to find doctors, lawyers, and dentists who were women.

As you'll see below, that all changed in the late 1960's and early 1970's.  What happened? It's probably a lot of things, but you could probably do worse than to point to birth control as a major contributing factor.

There are four views of the data from the Digest of Educational Statistics:

View 1 shows all degrees over time to men and women; the top via stacked bars, and the bottom using line charts.  The top chart shows the dramatic increase in degrees to women; the bottom shows that in 1955-56, almost all degrees (blue line) were awarded to men (purple line.)

View 2 shows the same data, presented a different way.  On the top chart, you see total degrees awarded, broken out by degree type.  Use the filter to limit the view to men or women.  On the bottom, degrees are awarded by percent of total: In the early 1960's, for instance, 99.62 percent of dental degrees were awarded to men.  By this decade, the totals had virtually evened out.

View 3 shows percentage change since 1955-56 for all degrees: Filter to law, dentistry, or medicine if  you wish.  Any way you look at the charts, the data are astonishing.  Especially interesting is dentistry, where there are actually fewer men graduating today than half a century ago.

And finally, View 4 shows institutions: There are now 70% more medical colleges, 60% more law schools, and 36% more dental schools than at the start of the analysis.  This latter number is interesting, however; while law schools and medical schools were at record numbers in 2013-2014, dental programs peaked in 1983-84.

Once again, women, given a more equal shot at education, outpace men by a considerable margin.

What do you see here? Leave a comment below.


Thursday, December 8, 2016

A Fresh Look at the New WICHE Data

Note: You should view this on a tablet or desktop.

The Western Interstate Commission for Higher Education (WICHE) has just published the 9th version of "Knocking on the Door," a look at demographic projections of high school graduates in the US.  And several organizations have already published interesting views of the data, like this one on the WICHE site and this one on Hechinger Report.

As I make my case for more self-service BI, these are great examples of what I call the 80/80 rule: Eighty percent of what an analyst will give you is not what you need as a practitioner; and 80% of your questions won't be answered when someone else does the analysis for you.  So I took the data (and allow me to complain a little bit about putting data for 50 states and DC in 51 worksheets in an Excel workbook, WICHE) and spent a lot of time restructuring it for analysis.  Then I started asking my questions, and came up with 6 views, in an attempt to provide practitioners maximum flexibility. On most of these, you can change start dates, end dates, states, regions, and ethnicity. Even after a couple of hours on this, I could have come up with 60 views, but my spare time is, of course, limited.

Overview:

I started with two high level questions.  The first is: What's going on at a macro level?  And thus this. The gray bars show total numbers of public high school graduates between 2000-01 and 2031-32; the lines break out those numbers by ethnicity.   So when your trustees ask about these numbers you can (but I wouldn't) say, "It looks like 2017 will be the low water point for at least eight years.  It's good news going forward."

Because if I were you, and before I talk to your trustees,  I'd want to know why lies underneath the macro trends, and thus my second question: How the changes look related to different ethnic groups: Take a look instead at the colored lines, showing break outs by ethnicity:  As I've written before, ethnicity matters because ethnicity and income and parental attainment all go together. The two groups that attend college at the greatest rates are showing shrinking populations (White students) or modest changes in numbers (Asians, growing at less than 20,000 by 2024.)

But wait.  Suppose you're in New York, or California, or Florida.  In all probability, your enrollment demographics are shaped more by what happens 500 miles from campus than 3,000. And Illinois? We've pretty much seen the best we're going to see until after I retire. So interact, and use those filters to get the view you want.

Tidbit: The population of White graduates peaked almost ten years ago.  And Hispanic populations will peak in 2024 after a long, steady, impressive increase since 2000.

View 1: National Overview


Issues of Control: 

It never fails: When I do the data on public high school graduates, someone asks about private schools. WICHE has included that data this year, and it's, well, sort of boring. The numbers are falling almost everywhere, and as a percent of the whole, are not keeping pace.  As you might expect, private school enrollments are a bigger thing in New England than elsewhere. (These data are not available by ethnicity.)

Tidbit: When a super-selective college brags that 60% of its students are from public schools, you can now understand that what they're really saying is private high school graduates are over-represented in their student body by a factor of five.

View 2: Public and Private



Digging down: The last four views (using the tabs across the top)

Now it gets a little more fun: Four ways to look at change over time.  In my business, knowing what's coming is very important; we have about an 18-year window on our markets, and no one's going to allow you to claim you were surprised.

I've become a fan of Hex Maps, to allow visualization of data on a choropleth map. It's never been easy to color-code on traditional state maps because Rhode Island and Delaware are so small, and Alaska and Hawaii are so far out there.

This map view allows you to see change over time (any two years you choose) by state; to show it for all public high school graduates, or just certain ethnicities, and to show numeric change or percent change.  I'm guessing you'll want to use this one a lot in strategic planning groups; it's my favorite.

Tidbit: Michigan. Notice how it stands out; that's a surprise to me.

Click to the next tab on the top, and you get the view for those who like bar charts.  You can still specify change between any two years, but this data is broken out by ethnicity. You can specify region or single state.

Tidbit: I was expecting bigger drops in New England, but no matter which years I focus on, the drop is smaller than I expected.  You?

The third tab is similar, but line charts over time.  It shows all ethnicities, allows  you to specify the years, and shows both numeric and percentage change from the first year chosen (be careful; these are not numbers, but rather changes).

Tidbit: Although Texas and California have seen dramatic increases in numbers of Hispanic graduates, it's Asian students in both states who will grow at the fastest rate between now and 2031.

Finally, the final view shows regions of the US, and the composition of the high school graduating class by ethnicity.

Tidbit: Pull the slider to show the changes over long periods.  Note how fast some regions (Far West, Southwest) change, and how slowly others don't (Plains, Rockies.)  Those last two regions will still be more White in 2032 than the whole US was in 2001.  And notice the relative stability of African-American high school graduates over time, as a percentage of the total.

Views 3-6: Changes over time

If you use Tableau, and  you want to download this workbook in its entirety, I can share it with you; I'll also be happy to send you the data in a much more accessible database format suitable for your own analysis.

I'd love to hear what you see on these visualizations or on your own analysis.  Leave a comment below.


Wednesday, November 16, 2016

Undergraduate Institutions of Doctoral Recipients, 2014

One of the most popular posts on this blog has been this one, where I showed the baccalaureate college of the nation's 2011 doctoral recipients.

This is an update to that, using 2014 data from the NSF.

It's pretty simple: There are three views here.

On the first view, you can see the undergraduate college of all doctoral recipients in 2014.  The view starts with known US institutions only, but you can add in foreign or unknown institutions if you'd like.  You can also look at a single state, the degrees awarded at the institution (for instance, if you have a student who really wants a liberal arts college, choose "Bachelor's-granting institutions"). Finally, choose the broad category of the doctorate, if you'd like, or even the specific program.  Note that the filters cascade: If you choose "Life Sciences" under Broad Category, you won't be able to find "Economics" under Specific Program.

On the second view (using tabs across the top), you can look at a single institution and find out how many graduates received a doctorate in 2014, by broad area.  Note that it doesn't matter if the person received the Bachelor's degree in 2010 or 1968: It's just everyone who earned a doctorate in 2014.

Finally, the last view shows institutions awarding doctorates, regardless of where the student originated.  You can see foreign and unknown institutions, US institutions, or all institutions as the baccalaureate college.

I hope this is helpful as counselors work with students on their plans.


Wednesday, September 28, 2016

Test score distributions, 2014

We tend to think a lot about a college's average test scores, despite the many ways colleges can and do manipulate them for their own benefit.  After my last post on the relatively low number of students who enroll in the most selective institutions, someone asked if I could do the same for test scores.  So here they are.

I've calculated very close mean ACT Composite and SAT CR+M means by taking the midpoint of the 25th and 75th percentiles.  They're almost certainly not perfectly accurate, but are very close, in all probability.  Then I've broken up enrollment to show where students attend college.

The first view is based on the earlier visualization; the second is a scatter showing both the ACT and SAT averages.  The first has just three filters; the second has more, plus a "Color By" parameter that allows you to color the colleges by one of several factors.

I hope this helps people think about and put score ranges in some context.

(Note: IPEDS does not collect test scores from test-optional colleges, or those that are open admissions.)


Monday, September 26, 2016

All the fuss, updated

One of the very first posts I did on this blog was showing just how many "Uber Selective" colleges and universities there are (or aren't), and how many students they enrolled (or didn't.)

I used it last week at a presentation at NACAC, and several people asked me if I had an update on it, so as soon as I got home, I pulled down the data and started visualizing it.  It's below, and it should be self-explanatory: Of the 1,943 four-year institutions shown, only 18 admit less than 13% of freshman applicants.  These institutions (blue bars) enroll just 82,000 students (under 15,000 of whom are African-American, Hispanic, or Native American), and only about 18,000 freshmen.  Yet they get a relatively large share of the press and attention whenever the discussion turns to college admission.

This has limited interactivity: You can choose region, public or private, or Carnegie group.

And of most importance: This is but a sliver of American higher education; for instance, 9% of all college students enrolled in the US attend a community college in California; and another 4% at community colleges in Texas.  Keep that in mind as you look at this data.


Tuesday, September 20, 2016

Who's Going to NACAC?

One of the things I hope to show people on this blog is that data is a lot more fun and interesting when you actually do something with it, rather than just present it in a spreadsheet. Here's a good example.

This week, over 6,000 people who work in or around college admissions will converge on Columbus, Ohio for the NACAC Conference.  (Yes, Oktoberfest is also in Columbus this weekend, and based on my informal discussions, there may be some overlap.)  NACAC puts its attendees in a table on its website for anyone to use.

But it's just data: What does a simple spreadsheet have the power to tell us?  Maybe more than you think.  Yesterday, I put the information in a visualization (first page is set up for mobile but autosized) designed to help people find other attendees.  As a side effort, I put up a chart of the most common first names of attendees, and it proved to be very popular. So last night I did a little more, and looked at most common first, and last names, as well as city, state, country, and organization.  They're below, and I think they say a lot about our profession.  What the information says is up to you to decide.

If you want to interact, click on a first name, and the other views update.  See? Interactivity can be fun too.

A note about the data: I did only minimal cleaning on it; when 6,000 people enter data on a form, there are bound to be errors.  Chicago, for instance, is not in Bosnia-Herzegovina. And I'm pretty sure Beijing is in China.  I did not clean up names, so if you really think your first name is "Mr. Daniel" you miss out on a chance to be included with the other Daniels. And Daniel is Daniel, not Dan, so variations are not grouped together.

Have fun.  And tell me what you think the data says.


Thursday, August 18, 2016

Tuition and Income in the States

Whoa, you might say as you look at this. It's way too funky for me. That's OK; I'm going to show you a new feature in the data visualization tool, Tableau, that I use that will make this all make sense. Hang on.

I wondered: Do states with higher median income levels charge more for tuition?  So I began to explore.

On each dashboard, median family income is displayed on the top chart, and college tuition on the bottom.  The view starts with four-year publics, but you can change it using the filter. The first dashboard shows only the rank of the states, from 1 to 5, with 1 being the high value in each.

If you can't make sense of it, don't worry: Use the little box in the upper right hand corner to select any single state, and that state's data will be instantly highlighted on both the income and the tuition chart.  You can see where a state stands on both measures.

The second dashboard (using the tabs across the top) shows the actual inflation-adjusted values (that is, $57,894 dollars in median family income, or $11,592 of tuition, both set to 2013), but the ranks are also displayed.  Use the state highlighter the same way, and hover over the dot for details. Note on this income chart I've broken one of my cardinal rules by not starting the y-axis at zero, for the sake of clarity.

You can get a sort of affordability index by looking at income ranks in comparison to tuition ranks, and you can see trends in both over time by state.

What do you notice here?



OK.  So maybe that's too funky.  Here's the same view, colored by red (high rank) to blue (low rank). If you like the original, it's below.


Wednesday, August 17, 2016

How Many Colleges Are There, Anyway?

A note in response to some questions from IPEDS geeks and others:  My data selection was from 2014 IPEDS data.  I used Title IV participating, US only, all sectors except administrative units.  That resulted in 7,018 institutions.  My visualization shows 6,876 because there were 142 institutions with absolutely no data reported.  I should have defined in my original post.

Also, the selectivity bands are not defined: Cut points are at less than 15%,, 25%, 40%, 60%, and 75%.  All others are "Not selective/Open."

College. University.  We think we know what these terms mean, and yet, any discussion of colleges in the US invariably leads to someone saying, "It depends on what you mean by college."

For instance, there are about 6,900 post-secondary institutions in the US, but only 2,654 offer a bachelor's degree; they enroll 10.5 million of the 17.6 million undergraduates.

Of all the institutions in the US, only 293 enroll at least 15,000 undergraduates, but this small fraction of colleges enrolls almost 40% of the undergraduates.  Conversely, there are over 4,300 options that enroll 1,000 students or fewer, but collectively they enroll only about one million students.  Our nation's public community colleges enroll over 6 million students on just over 1,000 campuses.

This visualization should give you plenty of options to see the shape of the higher education industry in the US: Filter and select to your heart's content, and as always, reset using the controls at the very bottom.

What surprised you?






Monday, June 20, 2016

Public University State Tuition

Note: The visualizations are not optimized for mobile.  A desktop is recommended for best viewing.

From the annual College Board Trends in College Pricing comes some interesting data, which I've combined into one database for visualization, focusing on public university tuition for residents and non-residents.  This looks complex, but it's pretty simple.

The opening view shows six charts: 2015 tuition for residents; for non-residents; and the premium a non-resident pays (in sticker price) across the top.  On bottom are three scatters: Resident tuition as a function of state funding per FTE student; five-year, inflation adjusted tuition for residents and not residents; and funding per $1000 of personal income and resident tuition.  Of these, I think the middle is the most compelling: Note the states that have raised tuition faster for residents than for non-residents.

The chart starts with US Averages in red, against the states as gray.  Use the control in the middle to highlight a single state on all six views.  As always, hover over any point for details, and use the reset arrow at lower left if you get stuck.

Using the tabs across the top, you can navigate to the map view.  Choose any value at top right to display on the map.  That value is displayed on the state, and the tiles (representing the states) are color-coded.  Red is high; blue is low.  Click on any tile on the map, and a summary of that state appears at the bottom.

Would your state legislator find this valuable? If so, I'd encourage you to forward to her or him. Otherwise, leave a comment at the bottom, letting me know what you see.



Wednesday, June 8, 2016

Public Institutions and Low-income students


Note: Visualizations are not mobile friendly.  I recommend a laptop or desktop for viewing this site.

Someone asked me today about what I thought higher education's biggest challenge was, and I said college costs without thinking.  And a few hours later, I still think that, with a twist: College costs for low-income students, especially at public institutions who presumably have a primary mission of educating students of all income levels in their state.

To be sure, costs are too high at private institutions, and many of the trends you'll see here are carried over and amplified in the private sector; but private colleges and universities may exist for different reasons, and that can be hard to capture in a visualization like this.

There are two views here, using the tabs across the top.  The first is a scattergram, arraying almost all 660 US, four-year public colleges and universities that admit freshmen (a few are missing data).  The x-axis shows in-state tuition in 2013, and the y-axis shows net price for freshman students who come from families with incomes of $30,000 or less, and who are paying the in-state tuition, most of whom are presumably in-state residents.  The color shows the percentage of students enrolled who receive a federal Pell grant, a program for very-low income students.

Reference lines show the unweighted, institutional averages, which allows the creation of quadrants, roughly:


  • The upper right, or high tuition, high net cost
  • The lower right, or high tuition, low net cost
  • The lower left, or low tuition and low net cost
  • The upper left, or low tuition, high net cost 

Color here is important: Red dots are those colleges with lower percentages of Pell students; blue dots show higher values, although I've capped the color range at 40%, about the national average, if you include all types of institutions.  It's important because it shows how many students these institutions enroll, not just how well they do at reducing price (if they do.)  In other words, it's a bit easier to do a lot to reduce cost for students if you don't do it for very many; it's harder on your budget if you enroll more.

You can limit the view to states, regions, Land Grant status, or by using the filters to show only institutions with certain admit rates or Pell percentages.  As always, take a look at California.  Well done, California.

The second view shows in-state tuition over time, accompanied by net price for three groups of students who receive aid.  Students from:

  • Families with income of less than $30,000 (gold)
  • Families with income of $30,000 to $48,000 (orange)
  • Families with income of over $110,000 (the highest band reported in IPEDS).  This is in blue.

The bottom chart on the second tab simply turns these numbers into an Net Cost: Tuition ratio.  A value of 1.5, for instance, means that the net price is 1.5 times tuition.  Note the definition of net price:  

Net cost shows all costs associated with cost of attendance, minus grant aid.  For example, a university may have a tuition of $5,000, but a cost of attendance of $17,000 to include housing, meals, transportation, and personal expenses.  If a student receives $10,000 in grant aid, that student's net price is $7,000, which is greater than tuition alone.

As always, hover for details, and use the reset button at lower left if you get stuck. 

What do you see here?  What else would you like to see?




Friday, May 20, 2016

Changes in In-State Freshman Enrollment in Public Universities, 2002-2012

This is a good example, I think, of how data visualization helps you make sense of things: Even simple things like a small table of data.

In this case, the table is from The College Board, showing changes in the percentage of in-state freshmen in our nation's public universities.  You can see the raw data by downloading Table 28, here. What you can't see by looking at that table, of course, is the overall pattern.  That's where a picture comes in.

There are only two numeric values in the table: Percentage of freshman enrollment that are state residents in 2002 and 2012.  I added a third, by subtracting one from the other.  Then I put them on a choropleth hex map, a format I like because all the states are the same size.  On this map, orange colors show states where the percentage of in-state residents has increased; purple shows a decrease, and grays are mostly even.

Be careful about interpreting this data. This visualization does NOT show, of course, that a university system is enrolling fewer in-state students; in fact, the number could have gone up if non-resident enrollment also increased, but at a faster pace.  It just shows what has happened to the makeup of that freshman enrollment: More in-state (orange) or less in-state (purple).

What do you see?



Monday, May 9, 2016

New SAT Concordance Tables

Note: I tweeted a link that was set up for mobile, and thus the visualization scrunched down to almost nothing.  If the URL has m=1 at the end, just delete it, or click on the title above to go to the desktop/iPad version.

The College Board just published long-awaited concordance tables to compare new SAT scores to old, and new SAT scores to ACT.

You can download the data here if you wish, or look at them visually below.  The tables in the data correspond to the tables on the visualization (that is, for instance, that Table 7 in the College Board worksheet can be viewed on Dashboard 7 here, using the tabs across the top.)

For convenience, Old SAT scores are always in light gray.  Notice also I've labeled the chart when the axes are not synchronized.

As this data is public, I have cited the original source, its purpose is educational, and this blog is not monetized, I believe the use of it in this format falls under Fair Use.

As always, hover over the dots for details.



Friday, April 8, 2016

A Deeper Dive on The Coalition Data

There has been a considerable amount of discussion in the admissions world about The Coalition for Access, Affordability, and Success.  I remain skeptical about the motives behind this, as I did when I wrote this in the Washington Post.  To be clear, however, I believe colleges have the right to create their own admissions platform and conduct the business side of higher education with great latitude. I am merely questioning how a fractured admissions process helps low income students find, apply, get admitted to, and enroll in college; and the use of the term "access" by colleges who have, in general, poor records of providing access to low-income students.

Many school counselors I've talked to are very concerned by what they perceive to be a dearth of information about how this all will work, and there are also lingering concerns about privacy, which have not yet been publicly answered (to the best of my knowledge), even though one component of the application platform--the Locker--is scheduled to open this month.

These colleges represent the very top of the pyramid among private institutions, and also include many large, state flagship public institutions, as well as a few statistical outliers.  But to look deeper at the data, I downloaded a large IPEDS data set, and just scratched the surface.  What should jump out at you is the impressive list of colleges, their collective wealth, and position on several of the scatter grams, below.

Use the tabs across the top.  Every view has a filter to show public/private/all institutions.  Coalition schools are in red to make them standout; everyone else is in gray.  The universe is about 1,945 four-year, degree-granting, Title IV participating colleges and universities in the Midwest.US.  (corrected 4/17 at 6:32 pm CST).

What do you see?



Wednesday, March 30, 2016

The Boom in International Enrollment

You hear a lot about enrollment of international students these days, and often, I think, when a subject gets a lot of play, it tends to be overhyped, often by people who don't really understand the data.

This would not be one of those times.

I used IPEDS trend analysis to look at enrollment of non-resident students (that is, students who are neither US citizens nor permanent residents) over time.  For comparison's sake, I also looked at overall enrollment over that same time.

This data set includes all 7,276 post-secondary institutions in the US, both degree-granting and non-degree-granting, whether or not they participate in Title IV programs, so my usual advice about IPEDS data is amplified a bit here.  Still, the trends are interesting.

The blue charts (on left) show total enrollment at these institutions: Bars show numbers, and the line shows percent change since Fall, 2004.  The red charts (right) show estimated international enrollment.  It's estimated because I had to calculate it using two variables, and the "percent of students who are non-resident" is expressed in a whole number, which is less precise than I'd like.

Of course, you're probably not interested in all the institutions in the US, so you can use the filters at right to look only at certain subsets, in any combination: Large doctoral universities in the west, for instance, or baccalaureate colleges in New England.

If you reset all those filters (reset button at lower left), you can look at any college or subset of colleges by typing the name in the box and make your selection(s).  If you get in trouble, just reset.

What interesting trends do you see here?




Thursday, March 17, 2016

International Enrollment and Engagement

The world is shrinking, if not literally, then metaphorically.  Some colleges and universities embrace this in big ways, and this is the purpose of this visualization.

The Institute of International Education puts out good data on both international enrollment at US colleges and enrollment of US students in study abroad programs.  I've combined that data into two views that show both.

The top chart contains two sort-able and filterable bar charts.  It starts out sorted from large to small on the left column, namely study abroad students in 2014-2015; if you'd rather sort by total international enrollment, hover over that x-axis until the small icon pops up and click that.  Reset by using the button at lower left.

The bottom charts shows every college in the data set, with study abroad on the x-axis and international enrollment on the y-axis.  Each dot is a college, color coded by control.

As always, if you want to look at a smaller set of colleges, use the filters on the right.  They will control both charts at the same time.

Three things: First, the data are for all students, graduate and undergraduate.  The IIE data are not broken out, so it's not possible to determine meaningful percentages, except of course for colleges that only enroll undergrads.  Second, the data is only reported for colleges that enrolled 10 international students and/or sent 10 students abroad.  Assuming the colleges reported the data.

So, on that point:  I've checked this data where I seen anomalies; there are several obvious colleges where it's missing.  A final caveat: High numbers can be caused by lots of things, including location, wealth of the student body, and curricular offerings, among others. There is (or should be) no value judgment attached to the numbers you find here.



Thursday, March 10, 2016

Election Results with Census Data

I normally focus on Higher Education data on this blog, and in fact, this visualization started out as a higher education post: I wanted to look at presidential election results from 2012 to see if education played a part in how people voted.  But since I had a large census file anyway, with lots of interesting information like income, ethnic groups, and other data, I decided to take it one step farther.  OK, may steps farther.  And to me, almost everything is ultimately about education.

If you don't like to interact with these visualizations, stop right now.  You'll have to play with this to see how it works.

On the top view, you see every county as a dot, color-coded by region, and arranged on a grid.  Hover over any dot for details, if you'd like.  Counties voting more heavily for Obama are on the right; Romney counties are on the left.  Wealthier counties are on top (higher median family income), and poorer are at the bottom.  Note the reference line at $53,046, the national median.

If you want to look at a specific state or region, you can do that using the filters.  But you can also look only at counties that meet certain demographic criteria, of your choice.

For instance, you could find counties that are at least 15% Hispanic and where at least 10% of the adults have a Bachelor's degree.  Once you apply the filters, only the counties that meet those criteria are displayed.  Use filters in an combination.  (Of course, you can't find any county that's 51% White and 51% African-American; the filters aren't magic.)

The data also shows up on the map at bottom; it's pretty self-explanatory: Each county is colored blue (Obama won) or orange (Romney won.)

As always, the reset button is at bottom.

I find this very interesting, and I hope you do too.  And I hope you vote in November; I need you for my next visualization!


Saturday, February 27, 2016

My homage to the Atlas of the Census

In eighth grade, a teacher introduced my class to the Statistical Abstract of the United States, and it changed the way I looked at numbers and information.  Of course, in those days, it was just a book, and very hard to interact with.  But I still remember poring over it, absorbing interesting, if arcane, facts about the country and its regions.  Ever since, I've been fascinated with reference books like it.

Then, sometime about 2005, I came across the Census Atlas of the United States, and my mind was blown even more.  It contains page after page of maps, representing gobs of data, visualized in a way that tells stories, sometimes with a single picture.  One map, especially, made an impression on me. It's in this chapter, page 14 of the chapter, page 40 of the book, showing Prevalent Asian Group, by county.  If you can't download that large .pdf file, it's below:


In my job, and in higher education in general, we think about race and ethnicity a lot, of course, but here was something that exploded and challenged the status quo and made me think differently: "Asian," was not a monolithic term.  Far from it.

That day, I bought a copy of the Atlas, and always pick it up when I see it.  It's astonishing, I think, and I still discover something new every time I spend some time with it.

I've wanted to do my own mini-version for a long time, but the thing that makes it hard is getting at Census Bureau data.  It's either very hard to extract, or rolled up in ways that don't inspire further analysis.  But I did come across a file that was accessible and fairly easy to work with, with data on a lot of factors at the US county level.  And I started working on it.

This includes 19 maps: The first 16 show US counties, colored by the variable indicated.  In the heading is the US average.  For instance, about 66.2% of the US population is "White alone."  All maps are color coded with that center value as the center point.  Check the legend.

You can look at the whole US or just a state; unfortunately, adding Alaska and Hawaii makes it hard and time consuming to get the maps to look good.  If you have Tableau, you can download this and look at it yourself.

The maps have data filters on them too.  So if  you're looking at the view with percent of adults with a high school diploma, for instance, you can pull the slider up or down to show only those counties with very high or very low values.  On each map, I've also added a somewhat-random second slider, so you can see combinations of variables.  On the high school graduates one, for instance, you can also look at population density.  See which counties have high/high, high/low, or low/low combinations. Play around.

Additionally, there are three other views: Two showing 2012 presidential election results, and one scatter-gram, where you can cross any two variables, and see the relationships between them.

If you work in my job, you might find interesting data that can help you understand how--and why--your distant markets are different than your local ones.  Or maybe it's just fun to play with.

Either way, I hope you enjoy it.  I'd love to hear what you discover.





Thursday, February 18, 2016

Educational Attainment in the States

In case the coverage of the 2016 presidential election didn't convince you, this might help you see why people in the rest of the country seem different somehow.

The data are from the US Census Bureau's American Community Survey, in 2012; this shows the educational attainment of adults 25-34 in that year.  Use the control at top right to pick a value to display on the map, and the states (represented by hex boxes here) change color to show the value you've chosen.  The bar chart also updates.

There are seven values:


  • Less than a HS Diploma shows the percentage of the population in that state aged 25-34 that did not complete high school
  • HS Diploma or Less shows the number above plus the percentage of people who have just a high school diploma
  • Exactly a HS Diploma shows just that: Everyone who graduated from high school but did not continue
  • HS Diploma or Higher is the percentage with at least a high school diploma, including everyone who went beyond that
  • Bachelor's Degree shows people who have just a BA or a BS
  • Bachelor's Degree or Higher and Graduate Degree should be self explanatory
Once you make a selection, the map and the bar chart update; both are color-coded with blue numbers lower and orange numbers higher.  Be careful with this: With low attainment rates, blue is presumably better (at least if you work in higher education); with higher attainment rates, orange is better.  Hover over a state to see the value, or look at the bar chart at the bottom, which displays the same data in a different format.

You may notice the map style; this is the first time I've used it, and I like it a lot.  It allows you to see values on small states that would otherwise get lost on traditional maps; and it allows Alaska and Hawaii to display just off the coasts without a lot of effort.  But I'd like to know what you think about them, too.