Motivate : a context-aware mobile application for physical activity promotion

The objective of this research is to explore the potential and feasibility of LBS on mobile platform in the field of healthy living promotion. The context-aware suggestion on physical activity intervenes at the right time and location. We are interested in the user acceptance of contextualized and personalized advices provided by a mobile application, and whether the users can be motivated to be more active. Therefore we aim at developing a mobile adviser application (Motivate) that provides context-aware physical activity advices to promote healthy living. Our endeavor began with a questionnaire study and a Wizard-of-Oz experiment to design the concept of a location and physical activity-based adviser system. Insights were gained concerning the user acceptance of the concept design and technologies. User location data, activity, geo information, weather, agenda, etc were proven to be important factors to understand the context and create motivating advices at the right location and time. The design and development of the Motivate application began with a web Service which generated advices based on user offline location data collected by GPS loggers. Information such as geo information, user agenda, profile and weather were processed with simple rules to find a suitable advice from the PA advice database. The feedback to the advices provided us with the confidence to proceed with implementing the real-time mobile application. The first version of the Motivate mobile application generated advices based on the user location sent by the mobile phone and integrated real-time inputs such as weather, agenda, geo information and profile. A small-scale user test was conducted to evaluate and improved this application. The features such as reporting actions and Reminder were added to the final version of Motivate for a better data collection and user interaction. The final user experiment was conducted to evaluate the final Motivate application. The results show that the Motivate application has the potential to change behavior and attitude towards an active life. In addition to the descriptive analysis of the final results, a binary logistic regression analysis was applied to predict models of user responses and self-reported actions, with respect to advice type, context and profile variables. Advice type and user context (especially time) are the most consistently associated with both the responses and actions. Some weaker associations of several profile variables with the responses and actions are shown as well. This research shows the potential of mobile applications that provide context-aware advices to stimulate behavior change and gathers real-time user feedback. It provides evidence of people’s willingness to use the new technology and the acceptance for real-time suggestions. The attempt and findings in this research also provide some design guidelines of how to incorporate context variables into different types of advices to promote active behavioral change