Trajectory Data Analysis using Complex Networks

A massive amount of data on moving object trajectories is available today. However, it is still a major challenge to process such information in order to explain moving ob-ject interactions, which could help in revealing non-trivial behavioral patterns. To that end, we consider a complex networks-based representation of trajectory data. Frequent encounters among moving objects (trajectory encounters) are used to create the network edges whereas nodes repre-sent trajectories. A real trajectory dataset of vehicles mov-ing within the City of Milan allows us to study the structure of vehicle interactions and validate our method. We create seven networks and compute the clustering coefficient, and the average shortest path length comparing them with those of the Erdős-Rényi model. Our analysis shows that all com-puted trajectory networks have the small world effect and the scale-free feature similar to the internet and biological networks. Finally, we discuss how these results could be interpreted in the light of the traffic application domain