Assessing the quality of Automatic Passenger Counter data for the analysis of mobility flows of local public transport systems

Automated Passenger Counters (APC) represent a valuable source of information for the analysis of mobility patterns on public transport systems. Despite data gathered through APCs ensure higher accuracy and completeness compared to traditional sources of data (e.g. surveys), these data may present errors that might hinder the capability of understanding travel behavior. Addressing the need to quantitively assess the quality of APC data, this study aims at creating a high-quality dataset for supporting the decision-making of public transport operators. The study presents the data validation phase for the identification of the major quality problems, including noise, missing data, and outliers. Furthermore, a comparison of data from APC with manually collected data on transit passengers is presented with the aim of quantitatively evaluating the accuracy of APC data.