Semi-Automated Processing of Harmonized Accelerometer and GPS Data in R: <i>AGPSR</i>

Abstract

The use of sensors in mobility, health, and place research is rapidly increasing, with the promise of objective quantification of human behavior patterns and the ability to understand complex patterns of human behavior in ways that were previously impossible. However, there is now a critical bottleneck in managing the resulting influx of sensor data. In particular, limited software pipelines exist that both clean and harmonize global positioning system (GPS) and accelerometer data into a single dataset that integrates information about time, space, and movement. Other approaches are costly due to proprietary software, personnel time, and being prone to human errors. To address this gap, we have created the open source AGPSR pipeline in the R software environment, combining GPS and accelerometer data to yield simultaneous information on location and activity status. Demonstrations of the pipeline and validity data are presented. The AGPSR software (Ver 1.0) is now available for researchers to use in studies involving simultaneous GPS and accelerometer wear.

Keywords

• sensor data
• software tools
• data processing
• data integration