The New York Times has this wonderful interactive feature today, where you can see where most cab pickups and dropoffs happen at any time of day on any day. It confirms a puzzling feature that I had already observed: getting a cab is hopeless at one corner, but if you move just one block over you are sure to get one. The map shows the number of cab pickups around NYU at 5pm on a Friday, the legendary time when it is most difficult to get a cab.
Most of the immediate NYU area (around Elmer Holmes Bobst Library) is a taxi desert, so you have to walk either west to Sixth Avenue (a well known hot spot along most of its length downtown), or east to Lafayette (for example, Astor Place). One thing that has always puzzled me is that it’s always very hard to get a cab on Broadway, running parallel to Lafayette just one block west.
Here’s one amateur theory: the less obvious hot spots (excluding train stations etc.) can emerge out of nothing. Over time taxi customers expect to get a cab on one street corner. Then taxis are more likely to cruise that street corner because that’s where the customers are. Both customers and taxis keep going to that street corner more and more as both sides come to expect the other to behave that way. And bang, you have now gotten the 1,425,674th example of spontaneous order.
This is a good metaphor for development because….
6 Comments
I think you make a good point – taxis probably go to where the business is. Maybe some of these hot spots randomly occurred over time and became reinforcing feedback loops?
It might also be worth while to consider:
- signals: how many intersections near the taxi pick up spots are attenuated? How many have a protected left turn? How many are near protected pedestrian crossings?
- Generators: what sorts of buildings / pop densities are found in the high pick up areas?
- Land use: in general what are the land uses in the area?
- Impact of transit – as you mentioned trains and bus stops could explain some of the trends (maybe taxis are hesitant to stop near a bus stop!)
Id be interested to see a pick ups/hour distribution and see when the peak hours are.
It’s a good metaphor for development because, assuming your guess is correct, there are many bizarre development policies that are in place which arose out of questionable beginnings! Development practitioners do make the people they are trying to serve walk a block east to get on their “cab” – be it conceptual or tangible – instead of going to the person they hope to serve.
…. it’s a “virtuous circle” which began when, for some reason, the expected rate of return to cruising that corner exceeded the cost, and a customer or two just happened to be there, which brought the cabbie back when more customers happended to be there, and away we go. Just like when a poor country entrepreneur has an idea, takes a chance, and just happens to stumble into a lucky situation that confirms his idea. So he does it again and, with some luck, is again successful. Now he has learned how the idea works, so he doesn’t need the luck anymore, and a successful business idea is born. Debonaire’s Pizza delivery in Kampala, Uganda comes to mind. No real residential street addresses in Kla., but the pizza is there in 60 minutes or less.
Geographic information system data like this New York taxi map enables evidence based policy. For example, decisions on road maintenance in poor countries can use GIS maps of levels of traffic and demand so that investment is directed to locations of highest economic rate of return, rather than to locations determined on political criteria. Budget decisions about funding of school and health facilities can similarly be based on GIS evidence rather than whim or sentiment or corruption. Data enables sound governance.
Great post, thanks.
You might be interested in a pair of recent posts on the Aid on the Edge of Chaos blog (www.aidontheedge.info) – one of which argues for politically savvy use of the ideas of spontaneous order and self-organisations (http://aidontheedge.info/2010/02/15/slime-mould-simple-rules-and-the-politics-of-self-organisation/) and the other which makes direct comparison between traffic management and development management (http://aidontheedge.info/2010/02/22/from-traffic-management-to-development-management/)…
The latter of these posts may be of special interest because it makes reference to the searcher-planner distinction in the White Man’s Burden, which clearly has relevance for the ideas you present here…
All best,
Ben
Spontaneous order? Unlikely. Broadway is slow and crowded with buses in the right lanes, which makes it hard to see potential fares. Lafayette doesn’t have buses and comes all the way up from Brooklyn Bridge, then becomes Park Ave South and goes to Grand Central Terminal.
Try 2nd Ave for downtown cabs.
Not a complete theory on it’s own.
Problem is, if one spot is a good taxi spot, then so is a spot 10 yards away. That means we should expect any spontaneous “good taxi spot” to migrate over a few years to a spot which is better for structural reasons.
It probably has more to do with proximity to major starting points and ease of access to routes to major destinations, the relative costs of walking a block or two, and the cost to the customer and the driver of the driver coming closer in to the true starting point (the driver pays to drive in empty, the customer pays to drive out again).
It would be interesting to analyse the differences between rainy days and dry days. I would expect that taxis spread out more on rainy days because passengers are less willing to walk a block.
3 Trackbacks
[...] This post was mentioned on Twitter by William Easterly. William Easterly said: Miracles of spontaneous order: cool map on how it's impossible to get a cab on one corner and easy one block east http://bit.ly/cuGlH4 [...]
Social comments and analytics for this post…
This post was mentioned on Twitter by mikegechter_rss: Miracles of spontaneous order: where to get a cab around NYU: The New York Times has this wonderful interactive fe… http://bit.ly/auUfoW...
[...] Easterly on spontaneous order on getting cabs in New York, and the relevance for [...]