Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.


To date, analysis of physical activity has focused on the use of a small number of simple summary statistics derived from uni-axial accelerometer data, such as the mean level of activity. These statistics only capture a small portion of the information in this data. Furthermore, higher-resolution tri-axial accelerometer data is now available in UK Biobank, in circa 100,000 participants. There is a need to develop methods to more comprehensively characterise patterns in these data, in order to investigate which aspects of activity relate to other traits and disease. In this talk I will present my initial work in this area, including a novel 'activity bigrams' approach that characterises how a person’s activity changes from one moment to the next. Activity bigrams can, for instance, differentiate between two people with the same time in moderate activity, where one person often stays in moderate activity from one moment to the next and the other does not.  We tested the association of activity bigrams with body mass index (BMI), as an exemplar, and identified several associations. For instance, a lower number of consecutive minutes in sedentary then moderate activity coupled with a higher number of consecutive minutes in moderate then vigorous activity was associated with a lower BMI, even after accounting for the amount of time spent in sedentary, low, moderate and vigorous activity overall.


Louise Millard is an interdisciplinary Data Scientist in the Integrative Epidemiology Unit (IEU) at the University of Bristol. Her research interests include the development of machine learning / data mining methods to automate epidemiological analyses and analyse epidemiological data. Louise is particularly interested in the analysis of time-series data, such as from accelerometers or continuous glucose monitors, to more comprehensively characterise patterns in these data.