I am approaching a data problem. My data set consists of observations of (X,Y) coordinates indicating a position on some grid. There are two groups based on a variable Z. Group A is all the points where Z>10 and Group B is all the points where Z<10.

Would it be a good idea to separately cluster the 2 groups of (X,Y) coordinates, then compare the location of clusters to reach some conclusion about how Z affects the location of the points?

I want to say something statistically about how differently, if at all, the group A and group B points are clustered. So I can say that the difference is because of Z.

  • $\begingroup$ I'm confused, is the value Z known for every point? If yes, why do you choose to do clustering? Overall is this more about visualizing/analyzing the data? $\endgroup$ – Erwan Oct 22 '19 at 15:22
  • $\begingroup$ The dataset includes X,Y,Z features. I am plotting X and Y. I want to see how a different value of Z will affect the behavior of the X and Y variables. Think of X and Y as pixels of an image. A different value of Z will affect what pixel is chosen. That being said, I don't believe regression would work here because I am not looking for how Z affects the values of X and Y. Rather, I am looking for how Z affects the cluster locations/cluster densities of the plot of X and Y. $\endgroup$ – Bran Oct 22 '19 at 21:32
  • $\begingroup$ Ah ok thanks, I thought it was the opposite (Z depending on X,Y). The first thing that comes to mind is to actually plot the points colored by Z and see how it looks like, but I guess you did that already? $\endgroup$ – Erwan Oct 22 '19 at 22:04
  • $\begingroup$ Yes, I'm just looking for something that says something about significance. So I can be sure that Z does affect the locations X,Y. (BTW: Sorry for any incoherence here, but this thread has really helped me gain clarity and confidence) $\endgroup$ – Bran Oct 23 '19 at 0:30
  • $\begingroup$ It's an interesting problem. I don't really know how to address it but I'm thinking maybe you could look at it this way: given a sample of points (X,Y,Z), does knowing Z helps predict other points more accurately? Not sure if it makes any sense though, it's just a thought $\endgroup$ – Erwan Oct 23 '19 at 11:55



In statistics, multivariate analysis of variance (MANOVA) is a procedure for comparing multivariate sample means. As a multivariate procedure, it is used when there are two or more dependent variables, and is typically followed by significance tests involving individual dependent variables separately. It helps to answer:

  • Do changes in the independent variable(s) have significant effects on the dependent variables?

  • What are the relationships among the dependent variables?

  • What are the relationships among the independent variables?

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.