Drafting the NBA: This Guy Built the Basketball Nerd's Infographic Wet Dream

In honor of the basketball season being extended by two more sweet days — thanks to perhaps the most incredible two minutes of basketball in history — spend a few minutes checking out this infographic by Adam Pearce, which measures a player’s efficiency rating against the order in which he was drafted. It really drives home the notion that most players, even those lucky enough to be drafted, never really contribute in a substantial way to a professional team.

The infographic is embedded below. Due to its frame size, it might be worth your time to check out the infographic on Pearce's site, before coming back to read his Q&A below.

This isn’t an effort to turn Motherboard into a sports page. From Obama’s data-driven re-election campaign to sabermetrics, we’re seeing the green shoots of the Age of Big Data in every field. And since we can record and store information about almost everything, mastery is less and less about collecting the most data, and more about presenting the data in a way that makes sense. The new gatekeeper is the graphmaker.

With a fresh degree in mathematics, Pearce has begun making interactive graphs in his spare time. He posts his results on his blog, Road to Larissa, along with his commentary and self-critique.

Making interactive graphs has got to be the most productive hobby I’ve ever heard of, so I recently caught up with Pearce to talk about his methods and goals, and try to figure out why he spent 40 hours proving that the 2003 NBA draft was utterly stacked.

Motherboard: How would you describe what you do on the blog?

I just wanted to have a blog. It seemed like a cool thing to have on the internet. I’ve just been putting up little projects I’ve been working on for the last few months.

Right now my focus is sort of interactive data visualization. I think it’s a really exciting time in that sort of space. Graphs have been around for hundreds of years and not really changed because there’s only so many ways you can draw a line graph or bar graph. With the development of modern web browsers, and the fact that everybody has one, you can start creating graphics that anybody, anywhere can interact with. I’m trying build things that facilitate that in a really intuitive manner where you don’t need any special training.

You’ve got a degree in mathematics, do you have an interest in design?

Not really. I wish I did, but I’m trying to read some books and get a handle on that.

Thus far, it seems like you're going with a really simple aesthetic.

Yeah, let me see if I can find this…

(Sound of Pearce clicking around on his computer)

I was talking with somebody about doing contract work—designing something for them—and they said, “Your visualitzations look good, and we like what you’ve produced. The only thing we’d add is that we want them to ideally look beautiful in terms of visual design. Would you be open to working with a designer?

When I’m working on something on my own, I’ve been aiming for a much more minimalistic look. I don’t have much for design brains so I feel safer being a sparse as possible.

You already have a pretty interesting range on your blog—from meteors to the NBA. When you’re working with data sets does it matter what it is? Do you think of it differently or is just data that needs to presented in a certain way?

When I first sit down, I have to get the data. At that point, I’m thinking a lot about what the data is. Is it in a spreadsheet somewhere where I can just upload it how it is? For the basketball statistics on Basketball Reference, I had to think about how I would download a webpage for every player and then restructure the data. So at that point I’m looking very carefully at what it is.

In terms of putting together the graphic, not really that much. And that’s something I should more aware of, because there are better and more specific stuff you can do if you know the domain better.

Like the meteor thing could be done with anything where you’ve got latitude and longitude coordinates.

How clearly can you picture the final product when you’re starting?

I have a little notebook that I keep with sketches and ideas. When I’m thinking of something to do, I’ll sketch it out. Generally [the final product] is pretty close. Sometimes I’ll see something really interesting that I want to show. I usually have a pretty good idea what I want it to look like and then I iterate on that, obviously, and make changes, but nothing too major usually.

So, with the NBA graph, it links back to Basketball Reference when you click on the individual players…

That’s where I got the data from, so it seemed good to link back to them. Also it was really easy to link back to them because I had already visited all those webpages and knew where all those players’ URLs were, as opposed to having to figure out Wikipedia pages.

Did you have to collect the information for each player yourself?

Exactly. I didn’t do it by hand, but I wrote a program to go to their draft page and download, like, the 1992 draft, and then for all the players whose names appear on that page, download all those player names and download their stats.

And that’s just called webscraping. It’s a pretty standard thing to do.

You mentioned and linked to The New York Times's football precedent. Is that what inspired you?

Exactly. You could probably go with shameless ripoff on my part, rather than inspiration. That’s why it ended up looking so similar to what I imagined, because it was somebody else’s work.

What other data are you looking at for the near future?

I saw a really cool map of tornados that did a pretty interesting thing. This was a data set with basically tornados, where they started and where they ended, their casualties, severity, and property damage. And the histogram, rather than being flat, was a curve — it went in a circle and showed the distribution of the direction the tornados traveled. And I thought that was super creative and I’d like to build on that in the same way that the meteor map worked to allow you to select certain filters. Like, “tornados that occurred at night,” or “hurt more people” or something like that. So take something that this person did and add more information to it, and make it interactive.

I remember seeing the map and feeling like it was weird that it wasn’t more interactive. It’s almost becoming the norm.

Yeah, and the sweet spot for my abilities right now is being able to implement that interactivity.