topic Re: correlation map with categorical and numerical factors in Discussions

correlation map with categorical and numerical factors

DBM_860752 — Tue, 29 Apr 2025 11:10:49 GMT

I have several dataset to analyze. I know how to do a color correlation map with numerical factors in JMP but I do not know how to do a correlation map with categorical and numerical factors. That would be very useful because correlation maps are a compact way to analyze and present a lot of information, especially when there are many factors.

How can I generate a correlation map with categorical and numerical factors?

Is there an alternative method to analyze correlations between many factors of different type?

Re: correlation map with categorical and numerical factors

Jed_Campbell — Tue, 29 Apr 2025 14:39:18 GMT

The Response Screening platform is really great at analyzing correlations between large numbers of factors, and it can compare lots of different data types. A good trick is to put all your columns in both the X and the Y role during the launch.

Re: correlation map with categorical and numerical factors

DBM_860752 — Tue, 29 Apr 2025 16:09:27 GMT

Thanks for the quick reply and for the suggestion. I have to go deep in the response screening since I am not familiar with that.

About the correlation map, I see that in the multivariate, it accepts only numerical values and the range for correlation coefficients is between -1 and 1. In the DOE module, I can evaluate a test matrix with a correlation map that accepts all factors, including categorical, and the range for the correlation coefficients is between 0 and 1. I am wondering whether this second type of correlation map is available outside the DOE module to make analysis of correlation of data from experiments.