The analysis of territorial changes over time raises many methodological questions. Despite a very frequent use of spatial analysis methods, few works underline their methodological shortcomings. This paper considers a comparison and discussion of exploratory statistical methods (notably Ripley's K function and its derivatives). It is worth noting three parameters in archaeological spatial analysis: scale, quality and quantity of input data. Archaeological data are intrinsically heterogeneous. Therefore, our approach proposes a multiscalar process integrating heterogeneous data. We describe a range of methods for the description and analysis of heterogeneous spatial point patterns in archaeology. These methods are described through the spatial analysis of neolithic sites in the South-East of the Paris Basin (France). We conclude by discussing the problems of the logical link between spatial pattern recognition and the archaeological process.