Modeling Trip-Length Distribution of Shopping Center Trips from GPS Data

Abstract

Automobile trip lengths are increasingly used in the calculation of development impact fees or bulk services contributions, affecting the revenue collected by local authorities. It is, however, difficult to obtain accurate estimates of current or predicted trip distances, and the empirical evidence base is relatively thin. Global positioning system (GPS) technology might provide a more accurate way of filling this data gap at a lower cost. The paper describes the use of mobile GPS loggers to collect and analyze trip-length data for car-based trips to and from shopping centers based on data collected from drivers in the Pretoria–Johannesburg area of South Africa. We verify the minimum stopped-time criterion used to identify trip ends under local conditions. Significant variation in trip lengths is observed, but average trip lengths vary systematically by shopping center type and size. Average GPS-derived trip lengths were found to be significantly longer than those estimated using conventional surveys in Florida, especially for smaller centers, raising the possibility that conventional methods lead to underestimation of the traffic impacts of individual centers. Although the study confirms the feasibility of using mobile GPS loggers for measuring trip lengths, several methodological questions remain to be solved to improve the robustness of the findings.