Using Insights and Incentives to End Rush Hour

A start-up called Urban Engines believes data analysis and commuter lotteries can help cities reduce congestion.

Image
Shutterstock.com

You can think of rush-hour traffic as a fairly straightforward problem of supply and demand. The supply is the capacity on city buses, trains, and roads. The demand is the share of metro area commuters who want to ride transit or drive cars. Congestion at the morning and evening peaks is a sign of demand overwhelming supply: too many commuters trying to use buses and trains and roads at the same time.

That's how the founders of Urban Engines, a congestion-relief start-up making its public launch today, prefer to view the city traffic problem. It's also the key to the company's proposed solution. On one hand, they believe new data insights can help cities adjust the supply of transit vehicles and road space. On the other, they plan to use behavioral incentives to control commuter demand.

"That's us in a nutshell: insights and incentives to attack congestion," says CEO and co-founder Shiva Shivakumar, a former Google engineer. "The better you can understand both sides of the [supply-demand] equation, the better you can start optimizing it."

On the insight side, Urban Engines relies on an approach called "crowd-sensing" to understand what's happening across an entire city transport system. Let's take the example of a subway. Each fare card entry swipe delivers basic information on rider location and (at least for cities that require a swipe in and out) total travel time. Using algorithms and supplemental data, such as real-time transit schedules, Urban Engines can deduce what's happening at any given subway station or train at any given time.

"Every single person in a crowd becomes a mini sensor," says co-founder Balaji Prabhakar, a Stanford professor of computer science. "Their overall trip plan actually tells us, when taken together, what's happening in the system."

That's a big improvement over the small platform or on-board samplings some cities currently use to get a sense of system flow. As a result, Urban Engines can produce interactive data visualizations that give short-term congestion insights (this platform is overcrowded, trains on this line are bunching) and longer-term traffic trends (on rainy days this station needs more cars). Transit operators can use that information to scheduled and dispatch train supply more efficiently.

Sample data visualizations from Urban Engines measuring traffic in a fictional city called Freedonia. (Urban Engines)

The incentives side of Urban Engines draws from programs that Prabhakar has helped conduct that pay commuters to travel at off-peak hours. Recognizing that too few cities had implemented congestion pricing plans, Prabhakar and collaborators have taken the opposite tack — rather than charge commuters who traveled during rush hour, they reward them for traveling outside it. In behavioral terms, it's a carrot instead of a stick, and Prabhakar says it's been successful so far in Bangalore, Singapore, and Palo Alto.

"Mostly anything that has a punitive quality, like a stick-style approach, seemingly just runs afoul of the working populous," he says.

In the original pilot study, held in Bangalore from October 2008 to April 2009, the incentives system worked incredibly well. Roughly 14,000 locals were given the chance to commute outside peak hours; every time they did, they improved their odds of winning a weekly raffle that paid out prizes ranging from $10 to $240. Over the course of the pilot, commuter traveling pre-rush hour doubled, and the average morning commute time of all bus riders fell from 71 to 54 minutes. 

Urban Engines, which has been developing its twin approach for two years, is already working with three cities. In São Paulo, the company is working to relieve the crowded bus system; in Singapore, it's using incentives to promote off-peak travel on the MRT railway; in Washington, D.C., it's in the early stages of data analysis for Metro. At the moment, Urban Engines is working only with transit agencies, though eventually they plan to develop ways to help commuters directly.

It's a bold response to a big problem, with some obvious obstacles before it. Most metro traffic emerges on roads, not on transit, which means Urban Engines will need to collect data through vehicle transponders (e.g. E-Z Pass) as effectively as it does via fare cards. (Fare cards themselves have an uncertain future.) Transit systems have limited equipment and personnel, and won't always be able to deploy more buses and trains even if they identify a need. Local employers play a role, too, since workers can't arrive early or late unless that's fine with the boss. Above all there's the generally entrenched nature of commute habits, especially among drivers.

Shivakumar and Prabhakar are aware of the challenges and ready to move forward despite them. "We're now at a stage where, from both the data side and the incentive side, we're ready to go from the thousands to the millions and onwards," says Prabhakar. Onwards and hopefully upwards.

Top image: Chatchai Kritsetsakul / Shutterstock.com

About the Author