Main statistical features for building a pedestrian mobility model

Abstract

We study the GPS data set of a collection of 260 students from up to 10 schools in the Barcelona Metropolitan Area on their trip from house to school. Participants collected the data with the help of a mobile application. After visualizing and cleaning the data, the main dynamical properties have been studied. We firstly study the difuusion of the movement and then we describe the instantaneous velocity with a log-Normal distribution, proposing a stochastic process to model it. We also obtain two typical stopping times for participants' stops duration. In addition, we calculate the orientation of the movement for each school and for the entire data set. Finally, we describe the change of orientation with a family of symmetric distributions on the circle, comparing the results for the different schools. We thus provide a first analysis to very valuable data to model pedestrian mobility with specific origin-destination journeys. Results found can also help to improve urban planing in the schools surroundings.