A researcher used big data from taxi meters to study decision-making of taxi drivers in New York. These data was well suited for this research.
One example of the simple power of counting the right thing comes from Henry Farber’s (2015) study of the behavior of New York City taxi drivers. Although this group might not sound inherently interesting it is a strategic research site for testing two competing theories in labor economics. For the purposes of Farber’s research, there are two important features about the work environment of taxi drivers: 1) their hourly wage fluctuates from day-to-day, based in part on factors like the weather and 2) the number of hours they work can fluctuate each day based on the driver’s decisions. These features lead to an interesting question about the relationship between hourly wages and hours worked. Neoclassical models in economics predict that taxi drivers would work more on days where they have higher hourly wages. Alternatively, models from behavioral economics predict exactly the opposite. If drivers set a particular income target—say $100 per day—and work until that target is met, then drivers would end up working fewer hours on days that they are earning more. For example, if you were a target earner, you might end up working 4 hours on a good day ($25 per hour) and 5 hours on a bad day ($20 per hour). So, do drivers work more hours on days with higher hourly wages (as predicted by the neoclassical models) or more hours on days with lower hourly wages (as predicted by behavioral economic models)?
To answer this question Farber obtained data on every taxi trip taken by New York City cabs from 2009 - 2013, data that are now publicly available. This data—which was collected by electronic meters that the city requires taxis to use—includes several pieces of information for each trip: start time, start location, end time, end location, fare, and tip (if the tip was paid with a credit card). In total, Farber’s data contained information on approximately 900 million trips taken during approximately 40 million shifts (a shift is roughly one day’s work for one driver). In fact, there was so much data, that Farber only used a random sample of it for his analysis. Using this taxi meter data, Farber found that most drivers work more on days when wages are higher, consistent with the neoclassical theory. In addition to this main finding, Farber was able to leverage the size of the data for a better understanding of heterogeneity and dynamics. Farber found that over time newer drivers gradually learn to work more hours on high wage days (e.g., they learn to behave as the neoclassical models predicts). And, new drivers who behave more like target earners are more likely to quit being a taxi driver. Both of these more subtle findings, which help explain the observed behavior of current drivers, were only possible because of the size of the dataset. They would have been impossible to detect in earlier studies that used paper trip sheets from a small number of taxi drivers over a short period of time (e.g., Camerer et al. (1997)).
Farber’s study was close to a best-case for a study using big data. First, the data were not non-representative because the city required drivers to use digital meters. And, the data were not incomplete because the data that was collected by the city was pretty close to the data that Farber would have collected if he had the choice (one difference is that Farber would have wanted data on total wages—fares plus tips—but the city data only included tips paid by credit card). The key to Farber’s research was combining a good question with good data. The data alone are not enough.