Temporal patterns of preferences through Wikipedia editing in different languages

Temporal editing patterns on Wikipedia provide a computational lens on cultural dynamics across linguistic communities. We analyse a decade of editorial activity (2001--2010) in eleven Wikipedia language editions and apply hierarchical clustering to two representations of collective behaviour: a static one, based on the thematic composition of each edition, and a temporal one, based on its weekly activity cycle within each thematic category. For the temporal representation we compare two normalisations: one retains only the shape of the weekly cycle, and the other combines it with thematic composition. The linguistic communities exhibit distinct circadian and weekly editing rhythms. Some groupings are recovered by both representations and by both activity metrics (edits and editors), most notably Russian and Chinese, whereas others, such as German and French, emerge only from temporal activity rhythms. Temporal information thus reveals structure that thematic composition alone does not, which points to time as a relevant and often neglected dimension in cross-cultural analyses of digital trace data, and in studies that use online platforms as proxies for collective cultural behaviour.