I found it very entertaining. And humanizing of the writer - it shouted "hey, look, this isn't your usual rant on punctuation" early on in the piece, before the writing style had had time to otherwise distinguish itself from the hundreds of other articles we've read on stuff like this. Like some weird kind of foreshadowing.
"Employment at graduation" is not a particularly meaningful metric, because STEM PhDs who are not employed at graduation usually just means that they haven't gotten a post-doc. They will take a few months and find a job in the tech industry.
In addition : The end-date of a PhD is indeterminate (at least in the UK), since one doesn't know for sure whether the thesis defense will require any re-writing, etc. So it's difficult to have a job fixed up completely on the day the magic letter arrives.
I chose to travel around Europe for a few months after school and then bothered looking for a job, so I guess by this metric I was unemployed.
In the end I left Kansas for a job in London, a much better choice for me than being employed at graduation.
Agreed - a bunch of pie charts (not great communicating info at the best of times) could hardly be called a "visualization" in the modern sense of creating a graphic that makes data easily understandable, visually appealing & useful
I would take pretty much anything with a temporal axis over these pie-charts.
Also I would favor stacked bars with optional trend lines over area charts in the case of countable annual data like this.
After which amount of data points is it considered recommended to switch to interpolated line/area charts?
