The characterization of the space/time traffic profiles in a cellular network can be of high interest for automating the operation of future networks, since the knowledge extracted from the traffic fluctuations in a cell and its neighbours can be effectively exploited by different optimisation functions. In this context, this paper takes as an input a set of real traffic measurements in a cellular network deployed in a large city and analyses, on a per cell basis, the traffic profile characteristics at different time scales (week, day, hour). Then, the analysis is extended to the space dimension by considering the traffic of one cell in relation to that of its neighbours. This allows identifying traffic complementarities between neighbour cells at different time scales that can be exploited by certain optimisation functions, as illustrated in the paper with specific examples.