Automatic recognition of "low-quality" vehicles and bus stops in bus services

A recent interest in transit service analysis resulted in advances in the monitoring of public transport quality from the passenger’s viewpoint. Several frameworks were proposed to show where and when different quality levels occur, but there has been no focus on determining which vehicles and bus stops lead to low-quality performance in bus services. This paper proposes a framework that: (i) performs a simple data collection on selected parameters on passenger activities at bus stops (e.g., consulting posted information) and in-vehicle (e.g., validating the ticket). This data collection is performed by Secret Shoppers on Origin Destination pairs representing paths travelled by passengers; (ii) proposes two new algorithms detecting criticalities for each route and parameter; and (iii) shows the vehicles and bus stops for which some targets are not met. These steps result in the first framework that can help build operational plans guiding the correction of criticalities arising in the delivered bus services. This framework is deeply investigated and discussed in a real-life Italian case.