When Chicago's Divvy bikeshare program was launched, a number of people asked: who'd use this? A hybrid of cycling and public transportation, it had to compete against both. Which it could do on cost, as a year of unlimited rides on Divvy is cheaper than a month of the same on the CTA, and cheaper than the cost of a typical bike or a typical bike tune-up.

But what about convenience? Recently Divvy held its second annual data visualization challenge, and one of the winners, by Shaun Jacobsen at Transitized, compares the speed of Divvy with the speed of the CTA. And Divvy wins by a nose.

Jacobsen's "Who's Faster" project starts with a look at the 1,000 top "station pairs"—i.e. the places that people most often go from point A to point B using Divvy. Then, those are compared to the same route on the CTA at noon on a Monday.

And a couple patterns emerge. One is that the bulk of station-to-station trips are faster, centering on five minutes' savings. It might not sound like much, but it adds up; Jacobsen calculates 32,023 hours saved over 571,634 trips. The other is that the most heavily-used station pairs tend to save more time than less frequently-used ones, as if people are starting to figure out how it works.

In my own experience, this aggregate reflects reality. I calculated the times for my own typical Divvy trips: 950 S. Wells and Union Station to the Tribune Tower.

For 950 S. Wells to the tower, it takes 25 minutes plus a three-minute wait, for a total of 28 minutes. Divvy takes 26 minutes, including one minute to get the bike and adjust it and a seven-minute walk from the station.

For Union Station—one of the most common starting and ending points for Divvy commuters—bikeshare takes 23 minutes. Taking the bus is 19 minutes plus a four-minute wait.

It's not a big savings of time for me, and Jacobsen's calculations bear that out. But it is competitive with the CTA, and much, much cheaper. (And if you have your own bike, it's faster still. The MIT Media Lab's You Are Here site calculates that, for trips of a couple miles, biking tends to be the fastest commute, even faster than driving if you factor in a buffer for parking.)

And it seems that people are figuring that out. Another one of the Divvy data challenge entrants, Abhinav Singh, did something of an annual report for the bikeshare program, finding that use grew considerably from 2013-2014, particularly among subscribers. Subscribers have gotten younger (from 37 to 35) and the percentage of female riders increased (from 21 to 25 percent). Here's what its growth looks like:

Patterns emerge here, too. Day-pass users are heavy early on—it launched in summer of 2013—and dominate the lakefront. But subscriber use picks up heavily away from the lake, as Chicago slowly figures out how to navigate the city by bikeshare.