A couple weeks ago, just in time for the Super Bowl, we published a couple fun graphics that used transcripts of ESPN’s “SportsCenter” as a way to look back on the NFL season.
Originally, I had a concept in mind very similar to the one we (mostly Shan Carter) did in 2010 for the World Cup.
A colleague suggested instead using 3D players rather than photos, in part just to do something new and in part to give us a way to put more players on a field at one time. Here’s a progression of sketches on that concept:
Original whiteboard sketch:
A drawing for how it would fit on a print page:
Graham Roberts’s proof of concept (with sizes semi-randomly assigned):
We added some labels and charts to Graham’s final rendering:
It ended up looking pretty cool and we were happy with it, but in the course of our analysis we really noticed a lot of funny quotes and cliches that the announcers said but I couldn’t really find an interesting way to present them.
We made a ton of charts looking for keywords we wanted to inspect, which let us sift through the data a little faster (though eventually we would have to weed out non-NFL references by hand). This output showed charts for mentions of words, both cumulative and week-by-week, along with a list of the usage of each word in context:
But presenting them was kind of a challenge. A straightforward approach (the only kind I know how to do, really) didn’t do much for anyone and took up a ton of space, so we dumped it:
We tried highlighting individual sentences (like “You’ll have more luck getting a ticktack out of the mouth of an alligator than getting information, especially about injuries, out of the mouth of Bill Belichick,” Aug. 10), but there wasn’t anything cohesive about a random list of quotes.
Then my boss said to write something original with the quotes as if I were writing for McSweeney’s. I said, great idea, imagining something like “Is It OK To Dunk On the President?”, one of my favorite McSweeney’s articles ever. Unfortunately, I couldn’t get it to work. Luckily, our intern, Ritchie King, who was already helping me with the analysis, was.
He turned a handful of silly cliches into a hilarious narrative about sports, war and Tim Tebow. We made his cliches piece the center of the graphic and had Sam Sifton read it online. (If you haven’t heard it yet, it’s worth a listen.)
Anyway, it was a fun project and proof that data is out there for almost any crazy idea. It also emphasized two important lessons. One, from Amanda Cox, is that you should make a hundred charts and pick the best one. We definitely did that – our project folder is full of boring analysis of various players and ideas. The second lesson is that the design and editing machine of the NYT graphics department can take a decent idea and turn it into something much better.
For the nerds out there, most of the analysis was done in R using the tm, openNLP and Rstem packages, but I can’t be sure which methods I used from which since Amanda just told me to import all of them.