An emergent strategy for characterizing urban hotspot dynamics via GPS data

The increasing volume of urban human mobility data arises unprecedented opportunities to monitor and understand city dynamics. Identifying events which do not conform to the expected patterns can enhance the awareness of decision makers for a variety of purposes, such as the management of social events or extreme weather situations [1]. For this purpose GPS-equipped vehicles provide huge amount of reliable data about urban dynamics, exhibiting correlation with human activities, events and city structure [2]. For example, in [3] the impact of a social event is evaluated by analyzing taxi traces data. Here, the authors model typical passenger flow in an area, in order to compute the probability that an event happens. Then, the event impact is measured by analyzing abnormal traffic flows in the area via Discrete Fourier Transform. In [4] GPS trajectories are mapped through an Interactive Voting-based Map Matching Algorithm. This mapping is used for off-line characterization of normal drivers’ behavior and real-time anomalies detection. Furthermore, the cause of the anomalies is found exploiting social network data. In [5] the authors employ a Multiscale Principal Component Analysis to analyze Taxi GPS data in order to detect traffic anomalies. The most of the methods in the literature can be grouped into four categories: distance-based, cluster-based, classification-based, and statistics-based [6]. Typically, due to the complexity of this kind of data, the modeling and comparison of their dynamics over time are hard to manage and parametrize [7]. In this paper, we present an innovative technique aimed to handle such complexity, providing a study of urban hotspot dynamics.