80

Spatial autocorrelation: lessons from scales.'” The use by geographers of local indicators of spatial autocorrelation has spread since 1995. Generally treated from a theoretical point of view, applications are still rare. However, the tools are now available to measure the spatial dimension of social phenomena. This paper proposes retracing the move from global to local indicators, not from a formal mathematical point of view, but emphasising a geographical reading of the results produced and what they say about spatial patterns. The aim here is to examine the potential and the limitations of these tools, and to encourage their use for data analysis in geography.