WiFi sensing for urban analytics: An open-source R toolkit

Abstract

WiFi sensing passively captures the WiFi packets that smartphones broadcast, offering continuous, low-cost pedestrian monitoring. The technique has been applied to footfall estimation, transit analysis, and public space studies. However, existing implementations remain tied to individual deployments, and the complete processing pipeline is rarely available as reusable code together with sample data. This paper presents an open-source R toolkit that processes raw WiFi data into five urban analytics metrics: Location, Count, Track, Revisits, and Activities. The toolkit spans the full pipeline from sensor assembly to metric extraction and is documented as an online book with executable code. Demonstrations on a university campus (24 sensors, 26 days) and in a commercial district (17 sensors, 9 days) show how the metrics describe temporal rhythms, origin–destination flows, observed visit frequency, and differences in inferred stays between street types. The five-metric framework offers a common foundation for pedestrian measurement that can carry over to other urban sensing technologies. All code and approximately 20.5 million pseudonymized WiFi observations from the two field deployments are archived in a publicly accessible repository, providing a methodological framework for future pedestrian sensing research even as MAC address randomization becomes increasingly widespread.