Ex-post validation of automated travel mode detection based on spatial-infrastructural and traffic-regulatory modal filters

Smartphone-based travel behavior surveys have become increasingly prevalent, providing detailed insights into travel behavior through automated travel mode detection. Despite advancements in machine learning and sensor-based classification methods, the validity and transparency of such data often remain unclear. The reason often lies in the fact that smartphone-based travel behavior recording relies on an external app provider, which processes and prepares the data. This paper introduces an ex-post validation framework for automated travel mode detection based on geo-referenced infrastructural and traffic-regulatory modal filters. The methodology is applied to data from the aspern Seestadt mobility panel, a longitudinal survey in a rapidly developing urban area of Vienna. By leveraging mono- and multi-modal filters derived from spatial and regulatory data, the study ensures reliable validation of travel mode detection accuracy. The results demonstrate that user involvement through confirmation and adjustment of detected travel modes can further enhance data quality. The proposed framework is straightforward to implement and transferable to other urban contexts, making it a valuable tool for improving data reliability in mobility research.