Last week David King, an urban planning professor at Columbia, posted to his blog a video that charts the origins and destinations of yellow cabs in New York City over a 24-hour period. The video, King explains by email, was made under his direction by graduate student Juan Francisco Saldarriaga, who animated a representative sample of 200,000 taxi trips. Think of it as the statistical equivalent to a New York City time-lapse. If that guy who narrates movie previews were here, he might begin, In a world where yellow cabs have mysteriously become red and blue dots… :
The cab activity itself doesn't strike one as exceptional. By 7 or 8 a.m. we begin to see lots of people taking cabs into midtown for work. At midday the activity in midtown starts to even out. Then at night the folks in midtown punch the metaphorical clock and hail a cab and scatter across the city, often toward the outer boroughs.
But King sees an important pattern for the data points: the origins and destinations have a geographical asymmetry that suggests people are only using cabs for one leg of their daily round trip. If this were a video of people driving their own car to and from work, the morning and evening rush would be a perfect mirror. It stands to reason, then, that the other leg of the trip is taken by public transportation; after all, it's unlikely that many people park their car somewhere then take a cab home.
In other words, writes King, New York City taxi cabs appear to work within the existing transit network, not against it:
This matters because it means that individual's travel journeys are multi-modal. If we want to have transit oriented cities we have to plan for high quality, door-to-door services that allow spontaneous one-way travel. Yet for all of the billions of dollars we have spent of fixed-route transit and the built environment we haven't spent any time thinking about how taxis (and related services) can help us reach our goals.
King, for one, has spent a lot of time thinking about this subject. He and colleagues Jonathan Peters and Matthew Daus of CUNY recently presented a paper on the complementary transit nature of taxi cabs at a meeting of the Transportation Research Board. In it, they argue that "taxi service is a critical aspect of a transit system."
There's a good bit of empirical evidence to support this position. In research from a few years back [PDF], Bruce Schaller, a former director of policy and the New York City Taxi and Limousine Commission, ran statistical models of cab activity in 118 U.S. cities and found three primary factors for taxi demand. Among them were the number of workers commuting by subway and the number of households that don't own cars. Both factors support the idea of taxis as part of a wider public transport network. (Interestingly, Schaller found neither population nor employment to be significant factors in taxi demand once his three primary factors were taken into account.)
There's also a good bit of common sense. Taxis enable car-less travelers to switch modes in the middle of a journey. A New Yorker can take the subway to work, walk to a bar, then cab it home, and many do just that every day. This "asymmetrical mode share," as King and company call it, is a hallmark of transit-oriented cities — granting easy, flexible travel to no-car residents:
The one-way travel of taxis allows people to use transit, share rides and otherwise travel without a car. In this way taxis act as a complement to these others modes and help discourage auto-ownership and use.
The asymmetrical pattern is not just citywide; it also holds true at the neighborhood level. The researchers looked at taxi trips in and around the Union Square part of Manhattan in February 2010. They found that the majority of the cab trips involving the neighborhood, nearly 57 percent, were origins. Many of these trips occurred in the late evening hours, which makes sense, since Union Square is home to New York University and thus a heavy nightlife. Presumably many people who reached the neighborhood earlier in the day by subway or on foot chose to hail a cab home after closing their tabs:
What doesn't add up is why many urban transport experts ignore the idea of using cabs to expand a transit network. The oversight is particularly vexing when you consider how much attention car-sharing receives as a means for reducing urban car ownership. Car-sharing is a great way to reduce ownership, of course, but the same might also be said for strategic reliance on taxi fleets, King and his colleagues argue. Cab networks even offer some benefits that car-sharing systems don't, such as a lower demand for parking.