Writing in the Harvard Business Review blog, Scott Berinato has interviewed top data visualizer Amanda Cox from the New York Times. Thanks to Nathan Yau for spotting this and posting it on Flowingdata.
Cox raises a couple of interesting points for me, points that are rarely said. Firstly, that statistics graduates leave university with almost no skills in creative problem solving and computing:
I come from a statistics background, and I’m finding statistics students’ portfolios are crazy weak compared to the computer science students, even though they’re playing with the same problems. I think it’s because comp sci students are encouraged to play, whereas stats majors it’s, “here’s your rule book, now make things.” I don’t think that’s the good model for making better visualization.
I think that’s absolutely true. I know because I am a stats (via maths) graduate myself, and everything I know about programming and visualization is self-taught in recent years. I mean no disrespect to my former teachers, it’s just that you can’t cover everything in the time available and the accepted norm is to teach the rule book. For the great majority of my fellow students, that’s exactly what they wanted: practical data analysis. But if you want to be able to do cutting-edge analyses, or create cutting-edge visualizations, you need different skills which are all about playing around with computers.
In fact, I’m working on a little example to post here soon, on how you can access the data behind such a visualization and play around with it to make your own version.