Posted on 07/06/2012 by


I am a busy person these days.  I have been teaching at a San Antonio community college and will be moving next semester to teaching High School Algebra I.  Also, among other things, I have been researching education data for a national nonprofit.

During this project, I needed to simplify some visualizations.  Not dumb down, but simplify.  And this reminded me of an important lesson: simplicity is powerful.

Now, it is easy to get caught up in cramming MORE information into smaller graphs.  While that is itself valuable (see for example my post on visualizing the U.S. Code), simplicity can also serve a purpose.  I have argued elsewhere that bar charts are the epitome of conveying information succinctly and powerfully.  In the report I finished recently, I found myself using lots of scatter charts and maps as well.

This shows a clear relationship between the number of households in a school district with Grandparents as primary caretakers and the 5-Year High School and GED Completion Rate.  For every percent increase in the prevalence of grandparent caretakers, the High School completion rate decreases by 2 percentage points.  To me, this was incredibly striking.  This was driven home by the geographic clustering of Grandparent caretakers.

I hope to share this and other similar data with policymakers as time goes on.  I think that such stark data can have a real effect by merit of there being nothing to distract you.  You can draw your own conclusions.  This is especially true with a graph like this:

For every roughly 1% increase in the number of households with more than one occupant per room,the number of students meeting TAKS (standardized test) standards decreases by 3%.  Whatever you may believe about nature vs. nurture or the effects of teachers or schools, this graph tells a story.  Home environment matters at school.

All of this just reminds me that hierarchical edge bundles can be cool, but a simple linear regression can–potentially–be transformational.

-Jacob Alonso