Spatial user-generated content (UGC) is increasingly being used to study a variety of geographical phenomena, including urban change in social and economic dimensions. Wikipedia content evolves over time and includes articles about geographical areas, points of interest, and geo-located events. In this article, we explore the spatio-temporal variability of geo-located Wikipedia pages, considering their complete editing history. Selecting Greater London as a case study, we study the association between Wikipedia activity and the socio-demographic characteristics of the spatial context. Editing activity grows rapidly at first, and is then followed by a slowdown, reaching a stable rate, with occasional spikes. The initial growth is distributed throughout the study area, but activity becomes gradually more concentrated in central areas. The socio-demographic variability is strongly related to the presence of Wikipedia pages, but only partially to the editing. This approach may support the detection and characterisation of socio-economic change at the urban scale.