Answers are grounded in this paper and cite their sources.
Background : Understanding human movement behavior is essential for transportation planning, public health, and health geography research. To conduct movement behavior research, researchers need efficient methods to process data sets reliably and ensure validity. Therefore, we present a validation analysis of an innovative algorithm for detecting participants’ trips and transportation modes, such as walking/running, bicycling, or vehicle. Material and Methods : A modular algorithm to obtain the participants’ trips and transportation modes was used and compared with an annotated data set of 40 adults who wore a Global Positioning System logger (Qstarz), accelerometer (ActiGraph), and wearable camera (SenseCam) for a mean of 4 days. Pictures from the SenseCam were annotated to classify whether each 2-min period was part of a trip and, if so, which transportation mode was used (walking/running, bicycling, or vehicle). Commonly used classification metrics were calculated for 15-s epochs. Results : The trip classification precision was 79%, with a sensitivity of 84% and an F1-score of 81%. The accuracy was 93%. Transportation mode classification performed best for bicycling (>90% for all metrics), followed by vehicles (>85%), and walking/running (>75%), with an accuracy of 90%. Conclusion : Our modular algorithm has good validity for detecting trips and identifying transportation modes in population studies. Its implementation is readily available as a Python package, making it easy for researchers to use.
No takes yet. Share an insight, caveat, or question.
A 2025 study studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: