There's a method called pattern matching, sort of clustering I believe. Could it be used to decide how to devide an area into regions? I have peatland initiation dates from a large area. I would like to find a way to define peatland initiation regions that may or may not be the same as some already known geographical regions. I cannot simply group the dates, of course; there's a variety of different dates in every region, but there are patterns in distribution of the dates that could be used in distinguishing different regions. Is it possible to cluster the dates, constraining the clusters geographically? Which programme could do this?