Manhattan traffic patterns
Manhattan traffic patterns | Wired Magazine June 2010

What works

The tendency with geographical data is to try to find a way to portray everything on a map. Surely, there is a map up there, and many people will recognize that the area is Manhattan instantly by looking at the map before they read it in the title. That’s a nice thing about maps – they transcend language and bad captioning to some degree. However, much of the detail is not to be found in the map. The map just shows us where congestion tends to occur, but it doesn’t tell us when we can expect these areas to be congested or just what “congested” means. In Manhattan, the average speed is under 10 mph so does congested mean less than 5 mph? Or what?

But if we look at the other graphs and charts it is a veritable jackpot of traffic information, at least at the collective level. I wouldn’t try to use this collection of information to plan your route through the city unless, of course, this collection of information causes you to take the subway instead of driving.

I hate pie graphs (as in the “Proportion of Miles Traveled”), but I am sympathetic to the triangulated pie graphs in the “Vehicle Distribution” graphic. At least it is visually easier to calculate the volume of a true triangle than a rounded off triangle. So if you find that you have to go with a pie graph, emulate the triangulated version found here and your viewers will come away with a better understanding of the information you are attempting to convey. I was surprised at how many people take taxis to get to work. But I am even more surprised at how many fewer trips there are on weekends. Fewer than half of those made on an average weekday.

Anecdotal evidence warning: When I first moved to Manhattan, I remember sitting in the car for two hours to drive around the block. There was a street fair nearby (not on any of the sides of the block traversed in this trip) and that seemed to slow everything to a standstill.

What needs work

I would have found a way to combine the average speed and the delays and associated costs. Clearly, the two are related – lower average speed must mean more delays. I had a little trouble understanding the delays and associated costs without the text from the article. If the speed and costs had been integrated into a single graphic instead of split into two (with a big pie graph in between), I think the link between speed, delays, and costs would have started to become more intuitive.

Here’s an excerpt from that section for the curious:

“In the end, Komanoff found that every car entering the CBD causes an average of 3.23 person-hours of delays. Multiply that by $39.53–a weighted average of vehicles’ time value within and outside the CBD–and it turns out that the average weekday vehicle journey costs other New Yorkers $128 in lost time.”

For more on how that was calculated, you’ll have to read the article. But the bottom line came down to a proposed $16 toll to enter Manhattan below 60th Street. It’s about what drivers in central London pay and the proceeds would go to bolster public transportation. Such an idea – known as congestion pricing – was proposed by the Bloomberg administration but voted down in 2008.


Salmon, Felix. (June 2010) “The Traffic Cop.” in Wired Magazine [infographic by Pitch Interactive].

Bonanos, Christopher. (17 December 2007) “Fare Enough” New York Magazine.