I'm using cluster analysis(hierarchical clustering) and visualizing each plot by Constellation Plot.
My questions are
1. When the dataset contains several groups, how to visualize the difference of each groups in one constellation?? For example, two groups can be visualized by colors??
2. When two dataset is visualized by each constellation plot mapping, is it possible to overlap two constellation plot mapping?? How to do it if it's possible?
Kiichi,
It's been a while I am trying to find the best way to answer your questions and I am still not fully convinced I know how to do it.
Your question 2 seems possible. I never tried to do it with constellation plots but this link shows how to make an overlay of two different plots. In short, you first need to customize each plot as you wish and then right click within one of the plots and choose Edit > Copy Frame Contents. Then, right click within the other plot and choose Edit > Paste Frame Contents. The order you do it will dictate which plot gets on top of which. It is crucial that you format each plot separately in a way that you want them on an overlay.
Ref to question 1, I am not so sure whether my answer here will address it the way you want but my take is the following. Ref to understanding the difference btw clusters, I'd use Red Triangle> Save Formula for Closest Cluster. This will create a new column in your table, its formula shows the rules for deciding which row belongs to which cluster, it calculates the distance of each row to each cluster, and allocates the rows to their nearest cluster by using the "If min" formula term.
I hope this answers your two questions to a decent extent, otherwise pls let me know. You may want to share the dataset your are talking about by anonymizing its column names, it should help me understand your questions further if needed.
