Identifying physical activity behaviours and population characteristics from secondary smartphone fitness apps

Oral Presentation B11.4

Authors

  • Francesca Pontin University of Leeds
  • Nikolas Lomax University of Leeds
  • Graham Clarke University of Leeds
  • Michelle Morris University of Leeds

DOI:

https://doi.org/10.14288/hfjc.v14i3.561

Keywords:

Big Data, Smartphone App, Measurement and Surveillance, Technology

Abstract

Background: Commercial fitness and activity tracking smartphone apps are becoming ever more ubiquitous and generating a large volume of physical activity data. Utilisation of this secondary data has the potential to provide new insights into patterns of activity behaviour. Purpose: To use commercial app data to identify temporal physical activity behaviour patterns and patterns of app usage, characterising the sociodemographic features of these behaviours. Methods: Daily activity data from 30,804 app users with 7 or more days of recorded activity over the course of 2016 were used. Sociodemographic variations in physical activity behaviour and probability of meeting physical activity guidelines were investigated. Moreover, unsupervised clustering methods were applied to activity behaviours to identify temporal patterns in activity behaviour over annual and weekly timescales. Results: We ascertain longer-term patterns of app usage with increasing age and more male users reaching physical activity guideline recommendations despite longer daily activity duration recorded by female users. Key seasonal and weekly patterns of activity behaviour were identified within the app user’s data. Daylight saving was shown to play a key role in influencing activity behaviour across the clusters, with increased activity in summer months. Investigation into weekly behaviours identified varied roles of weekday versus weekend on activity levels. With some users being more active in the week and some on weekends. Conclusions: Fitness and activity tracking apps are valuable as data sources of activity over long temporal periods, allowing identification of patterns and characteristic of different socioeconomic groups, not often seen in shorter duration studies. Funding: ESRC: Data Analytics and Society Centre for Doctoral Training (ES/R501062/1).

Published

2021-09-30

How to Cite

Pontin, F., Lomax, N., Clarke, G., & Morris, M. (2021). Identifying physical activity behaviours and population characteristics from secondary smartphone fitness apps: Oral Presentation B11.4. The Health & Fitness Journal of Canada, 14(3). https://doi.org/10.14288/hfjc.v14i3.561