With the emergence of mobile phone trace datasets, new opportunities have arisen for improving the understanding large-scale mobility behaviours. The potential impact of these insight derived from these data is no more significant than in the developing country context, where existing data collection infrastructure is limited or non-existent. In this research, mobile phone data for Dakar, Senegal is used to better understand urban activity and mobility dynamics. To achieve this, a clustering method is introduced that extracts the spatial distribution, and the temporal characteristics, of the activities of individual mobile phone users. With this classification of individual locations of activity, citywide trends in activity and mobility over time are derived. The paper concludes in discussing the potential and limitations of this approach, and the outlook for associated analyses that employ mobile phone trace data.