The pNEUMA dataset is the result of a unique field experiment using a swarm of ten drones flying over the central district of Athens, Greece. Using this newly available dataset, this paper focuses on two directly related topics: i) lane changing and ii) lane choice in one of the busiest arterials in the city (Panepistimiou avenue). The main concept of this study relies on the definition of two layers regarding how the six lanes of the arterial are actually being used. Specifically, we argue that the lanes that are marked on the arterial (marked layer) are influenced by the frictions created by buses, bus stops, taxi stops, illegal on-street parking, etc. Using a lane detection algorithm, we show that there is an active layer which is different to the marked lanes on the road affecting capacity in the macroscopic level and driving behavior in the microscopic