Aggregated versus individual passenger traveling patterns in urban metro system

Data of the automatic fare collection system of the city Shenzhen in China was analysed to extract passenger traveling distance and the corresponding dwelling time for each passenger in the metro system, so as to compare passenger trip patterns at both aggregated and individual level. It is found that the traveling distance pattern transit from exponential decaying at short distances to power-law scaling decaying at long distances. Further analysis indicates the passenger volume plays an important role in affecting the dwelling time, which forms the special distance distribution. At individual level, passenger trip features, including the day-to-day traveling distance distribution, as well as inter-travel time distribution have been investigated. Results indicated that although passenger entering and exiting the metro station show a time-clusterized behaviour at aggregated level, we can hardly find any universal pattern for individual trips. Further quantifying a relatively long-time traveling features of the passengers, we can see that some of the passengers are characterized by a regular or quasi-periodic behaviour in time; some others are characterized by Poissonian behaviour.