The increasing urbanization in the last decades results in significant growth in urban traffic congestion around the world. This leads to enormous time people spent on roads and thus significant money waste and air pollution. Here, we present a novel methodology for identification, cost evaluation, and thus, prioritization of congestion origins, i.e., their bottlenecks. The presented work is based on network analysis of the entire road network from a global point of view. We identify and prioritize traffic bottlenecks based on big data of traffic speed retrieved in near-real-time. Our approach highlights the bottlenecks that have the most significant effect on the global urban traffic flow. We follow the evolution of every traffic congestion in the entire urban network and rank all the congestions, based on the cost they cause (in Vehicle Hours units). We show that the macro-stability that represents the seeming regularity of traffic load both in time and space, overshadows the existence of meso-dynamics, where the bottlenecks that create these congestions usually do not reappear on different days or hours. Thus, our method enables to identify in near-real-time both recurrent and nonrecurrent congestions and their sources.
© 2022. The Author(s).