David Zipper is a Visiting Fellow at the Harvard Kennedy School's Taubman Center for State and Local Government. He writes frequently about the future of urban mobility and technology.
If ride-hailing companies want to act like public buses, cities will need their numbers to make policy decisions.
The ride-hailing service Lyft recently attracted headlines—and some ridicule—when it launched Lyft Shuttle, a service in San Francisco with fixed routes and pickup locations. As many noticed, this bold new high-tech mobility innovation looks remarkably like a city bus service.
Lyft Shuttle is only the newest private option to get around town; commuters in tech-savvy cities like San Francisco can choose from a growing array of services offered by Lyft and Uber, as well as shuttles like Ford’s Chariot and GM’s Maven, and vehicle rental platforms like Zipcar and Getaround. Private “rogue” bike-share services are also in the offing for many U.S. cities.
The emergence of these mobility options is good for commuter choice—provided, of course, that society protects the traditional role of public transit. (Noting how Lyft Shuttle in San Francisco strategically avoids low-income areas, Salon warns that private bus-like services will exclude less-affluent riders, furthering transit inequities.) But there’s another, hidden problem: When a passenger decides to take Lyft Shuttle instead of MUNI, she is moving more than her money from public transit to a private company—she is also moving data about her trip. But no private transportation service today provides local government with point-to-point information about their passengers’ rides. The result is transportation policy that cannot be optimized to serve residents.
Let’s pause for a moment to consider why private data is such a critical part of transportation policy. Imagine you’re the director of your city’s transportation agency, and you’re looking for a way to ease congestion in the evenings at an intersection near bars. Ride-hailing companies have suggested that you take out a parking meter and turn the street space into a pickup/dropoff zone, predicting that drivers will spend less time circling the block trying to find their passengers (and thereby contributing to congestion). That sounds great, but your city gets $10,000 annually from that parking meter. Should you do it?
If you’re a good technocrat, you’d ask for data before making a decision. How many passengers are getting picked up at that intersection now? Within nearby blocks? How much traffic on a weekend night is rideshare vehicles, and is there evidence that they are circling trying to find their passenger? Without the trip data from Uber, Lyft, et al, it’s hard to tell. But that’s precisely the position today’s planners and transportation executives are in. (Note: Uber makes some ride information available through Uber Movement, but only by neighborhood and average trip time, not specific locations or individual rides). Such data limitations can constrain other policy decisions too, such as implementing road diets or closing lanes during construction. The National Association of City Transportation Officials (NACTO) has published data sharing principles for private transportation companies, but current practices fall far short.
There are several reasonable arguments to explain why these companies can’t simply hand over individual trip data to the public sector. Issues of passenger privacy are at stake, and valuable business information could be leaked to competitors. There are simply too many one-off data requests from local government for the private companies to comply with each one—and government staff may not know how to analyze and understand the data even if they have access to it. Each of these arguments has merit. But as private services like Lyft Shuttle develop and scale, city officials are forced to base policy decisions on data derived from public sources, and that information is becoming less representative of residents’ total trips.
Is there a path forward?
The answer might be found in a very unrelated field: cardiology. Cardiovascular disease accounts for one of every six dollars spent on American healthcare, and hospitals compete fiercely to maximize their share. But there are only so many patients who will receive care in a given hospital, which limits the data analysis possible in any hospital (or network of hospitals) that wants to improve treatment. An obvious solution would be for hospitals to pool their patient data so that researchers can answer questions like “is drug A or B better for a patient with XYZ conditions.” But competitive pressures among hospitals stack the deck against such peer-to-peer sharing.
Twenty years ago, the American College of Cardiology (ACC) proposed a solution. ACC would act as a neutral broker, pooling anonymized patient information from hospitals nationwide through their National Cardiovascular Data Registry (NCDR). NCDR’s data would be available for analysis at carefully selected analytic centers (universities like Duke and Yale), with groups like drug companies paying for the privilege to access it.
“We’re neutral among patients, payers, providers, and pharmaceutical companies,” says Kevin Fitzpatrick, formerly the chief innovation officer of ACC, now CEO of the American Society of Oncology’s new data registry, CancerLinQ. “The ultimate goal of these big data initiatives is to allow a professional association to serve as an honest broker of the evidence—ideally becoming, if you will, the single source of truth.”
Here’s how that model could be applied to urban transportation. All providers of transportation services—public transit agencies as well as private companies—would hand over anonymized trip data to a trusted third party that would standardize it and ensure trade secrets are not revealed. Researchers, city officials, and company employees could then access the data through a limited number of analytic centers—likely universities with strong transportation or urban planning programs. These centers would ensure neutrality and help registered users utilize the data to test their hypotheses.
To return to the earlier example, a DOT official could run the numbers to test how much congestion would be alleviated by converting a parking meter to a pickup/dropoff location. The transportation companies could pay for the opportunity to run their own analyses as well—and they would no longer have to worry about one-off requests for information from city officials who may need hand-holding to do analysis. And the public would benefit from more data-driven policy.
A data registry is not easy to establish, and it will not happen overnight. But if the rapid ascent of shared ride-hailing shows us anything, it’s that private mobility providers are poised to handle a bigger share of total urban trips in the future. If we want government decisions to be based on hard numbers rather than hunches, we need that data, and we need it ASAP.