Leveraging IoT Lifelog Data to Analyse Performance of Physical Activities

Internet of Things (IoT) based devices and wearable sensors have become ubiquitous accessories in our lives. Access to digital devices such as a smartphone, smartwatch, tablet, laptop or digital camera is continuously growing as consumer electronics and network coverage increase. Our day-to-day activities can be monitored and recorded using IoT devices. This data capture is referred to as lifelogging. For example, a smartwatch can track and record various information about our physical activities (e.g., step counts) and health status (e.g., heart rate and stress level). Recording and analysing lifelog data paves a way to investigate and expand the understanding of people’s life and their experience. This paper utilised a publicly available lifelog dataset of 14 individuals to understand how subjective wellness attributes (e.g., mood, fatigue, readiness and soreness) can influence their exercise performance (e.g., weight loss, burnt calories and exercise speed). The result of this study shows that the signals captured by the consumers’ electronics are sensitive enough and sufficient to understand the association between subjective wellness factors and physiological signals. It is illustrated that the subjective wellness attributes influence the sensors recorded exercise performance. However, the strength of wellness association on exercise performance depends on the exercise metric and type. The findings of this study help exercise recommender applications to take the individuals wellness attributes into account and develop a personalised exercise recommender systems.