I have a longitudinal dataset of (c.a.) 100 observations. Each observation is assigned to a point in space.
In each period, each observation is assigned to one out of x groups (i.e. no multiple memberships), where x changes every period. As a result, I have x groups distributed on a map, whose number and members change in each period.
I want to know if, on average, the spatial distance between the members of the groups decreases over time: e.g., groups become more "cohesive" in the map.
In order to answer to this question, my idea was to find a sort of index of spatial clustering for the groups in one year, and then observe its trend over time. But I cannot find an index considering multiple groups at the same time.
Does anyone have any suggestion or reference?
