Project OverviewMy PhD project, “Taxonomy of Movement Behaviours,” focuses on enriching movement data with greater contextual detail. I develop and validate novel methods that integrate data from accelerometers and GPS sensors to infer not just if a person is moving, but also how, where, and why. These techniques can identify transportation modes, characterize behaviours at specific locations, and uncover daily patterns. The ultimate goal is to deepen our understanding of human behaviour, enabling the creation of personalized interventions that foster healthier lifestyles.First StudyUnveiling Trips: This project introduces and validates a method that utilizes collected GPS data, contextualizing it by identifying an individual’s trips and stationary periods. It further identifies the mode of transportation—such as walking, bicycling, or vehicle use—to quantify daily travel behaviour, including the frequency and duration of active transport. This approach effectively uncovers how individuals travel throughout their day.Second StudyDeveloping and Validating Human Activity Recognition Algorithm: This work presents a robust validation of ActiMotus, a new open-source Python algorithm for Human Activity Recognition (HAR). Optimized for large-scale studies using thigh-worn accelerometer data, the algorithm’s robustness and generalizability were tested across multiple datasets. This validation spanned two device brands, three distinct age groups, and varied sensor configurations to ensure its broad applicability.Third StudyUnveiling Locations: This project introduces and validates a method that contextualizes GPS data by identifying locations where individuals spend time. Leveraging open-source GIS databases, the method classifies these places into domains such as work, school, leisure, and home. This enables exploratory analysis of behaviour patterns and provides insights into the time people spend within these key locations. |