Understanding routing behavior is an important task in modern transportation engineering. Effective management of routing behavior can reduce the travel demand on the street bottlenecks and thus help to relieve traffic congestion. Nowadays, there is not a standard and large scale data set collected over long periods of time that allows us to characterize route choice behavior.
In this work, we present a method to derive the routes taken passively from location based services data, the byproduct of the use of smartphone applications. The method is effective at inferring the routes from traces with varying sample rate intervals. Taking the Dallas-FortWorth metroplex as an example, we relate the route choice behavior to the trip characteristics and traffic congestion, confirming that travelers taking alternatives could reduce their travel time and increase the reliability during peak hours. Further, by integrating the route choice behavior with the imputed socio-economic characteristics of travelers, we find that the differences in route choice behavior of people by income levels is insignificant in the study region. The proposed data analysis framework is cost-effective to treat sparse data generated from the use of smartphones to inform route behavior. The potential in practice is to inform demand management strategies, by targeting individual users while generating large scale estimates.