The 2019 Active Living Conference will be held February 17-20, 2019 at the Francis Marion Hotel in the heart of historic...
Identifying Walking Trips Using GPS Data
Cho, G., Rodríguez, D.A., & Evenson, K.R. (2011). Identifying Walking Trips Using GPS Data. Medicine and Science in Sports and Exercise, 43(2), 365-372.
PURPOSE: This study developed and tested algorithms to identify outdoor walking trips from portable global positioning system (GPS) units in free-living conditions. METHODS: The study included a calibration and a validation phase. For the calibration phase, we determined the best algorithm from 35 person-days of data. Measures of agreement regarding the daily number and duration of diary-reported and GPS-identified trips were used. In the validation phase, the best algorithm was applied to an additional and separate 136 person-days of diary and GPS data. RESULTS: The preferred algorithm in the calibration phase, resulted in 90% of trips identified from the GPS data being found in the diary, whereas 81% of trips reported in the diary being found in the GPS data. The preferred algorithm used (1) a maximum 3-minute gap between points to define a trip, (2) at least five or more minutes of continuous GPS points, (3) a speed range between 2 km/h and 8.0 km/h, (4) at least 30m of displacement between the start and end points of a trip, and (5) merged walking trips when the time gap between trips was less than 3 minutes. With the validation data, substantial agreement between the GPS and the diary was achieved, with 86% of trips identified from the GPS data found in the diary and 77% of trips reported in the diary found in the GPS data. CONCLUSION: The algorithm identified free-living walking trips of more than 5 minutes in duration. The ability to identify outdoor walking trips from GPS data can be improved by reducing recording intervals used in the GPS units and monitoring participant compliance. Further research is desirable to determine whether concurrent wearing an accelerometer may improve the ability to detect walking more accurately.