Uncovering cabdrivers’ behavior patterns from their digital traces

Recognizing high-level human behavior and decisions from their digital traces are critical issues in pervasive computing systems. In this paper, we develop a novel methodology to reveal cabdrivers’ operation patterns by analyzing their continuous digital traces. For the first time, we systematically study large scale cabdrivers’ behavior in a real and complex city context through their daily digital traces. We identify a set of valuable features, which are simple and effective to classify cabdrivers, delineate cabdrivers’ operation patterns and compare the different cabdrivers’ behavior. The methodology and steps could spatially and temporally quantify, visualize, and examine different cabdrivers’ operation patterns. Drivers were categorized into top drivers and ordinary drivers by their daily income. We use the daily operations of 3000 cabdrivers in over 48 million of trips and 240 million kilometers to uncover: (1) spatial selection behavior, (2) context-aware spatio-temporal operation behavior, (3) route choice behavior, and (4) operation tactics. Though we focused on cabdriver operation patterns analysis from their digital traces, the methodology is a general empirical and analytical methodology for any GPS-like trace analysis. Our work demonstrates the great potential to utilize the massive pervasive data sets to understand human behavior and high-level intelligence.

Research highlights
► In this paper we developed a novel methodology to understand cabdrivers’ operation patterns by analyzing their continuous digital traces. For the first time, we systematically study large scale cabdrivers’ behavior in a real and complex urban context (3000 taxies in a metropolitan area). We identify a set of valuable features, such as RRSL and RRST, which are simple and effective to classify cabdrivers, delineate cabdrivers’ operation patterns and compare the different cabdrivers’ behavior. We thought that top cabdrivers would care about length of trip; instead they care about the time of trip, and they want to maximize their usage of time and finish their tasks as soon as possible. Thus top drivers usually have higher speed than ordinary drivers both on operation status and idle status. Because top drivers have better knowledge of street networks and traffic condition, they have more flexibility during the longer trip.
► We look forward to real time optimization and feedback: Could the overall spatial–temporal demand data be shared with all taxis so that they can better cover demand areas? Could everyone benefit from the intelligence of the top driver, following him in order to increase their revenue? Could we design a better system for providing real time feedback to individuals to optimize system performance? This is where pervasive computing can allow us not only to better understand our cities, but also to make them better.
► What does this analysis have to do with the city? What does it mean for the city? This analysis is interesting from a mode choice and taxi demand point of view. It is also interesting for city planners because it may dispel notions about where taxi drivers find the highest need. An expected guess would be that taxis find their most profit in the Central Business District, but the observations showed that high-earning taxis made their profits in another part of the city. This may come as a surprise to planners, who may react by building taxi stands and safe standing areas, re-routing busses, or posting taxi company phone numbers for potential patrons.
► It is also interesting to infer activities from the temporal resolution of the traces. When do conference-goers need a ride from the airport to a hotel? Do they arrive on Monday mornings and stay for three days? What does this mean for convention center placement? Which attractions are being visited by tourists via cab? Do people seem to be choosing taxi service for their daily ride to work? Are nightlife institutions producing a significant amount of nighttime calls for rides home? These questions are helpful for a city because they describe resource and space usage.
► These findings are inspiring because we have various assumptions about human behavior in the social life but seldom get the direct data sets to prove or defend them. The study about cabdrivers’ operation behavior utilizing the massive GPS traces opens the new venue for the research about human behavior. We believe this type of analysis and methodology is replicable and well-suited for a number of application domains like logistics, mass transit, pedestrian initiatives, or zoning. We are excited to see that pervasive data sets can help us understand human behavior and motivational intelligence.