[R] MFA variables graph, filtered by separate.analyses

gavin duley gdu|ey @end|ng |rom gm@||@com
Tue Feb 21 16:24:19 CET 2023


Apologies if this is not the correct place to ask. I am attempting a
MFA analysis of a dataset based on wine chemical and sensory analysis,
based on the STHDA tutorial [1]. (I am using this dataset here too, as
an example dataset to work on without posting my actual data. I've
tried this with both my data and the example data, with the exact same

The only issue I am having is that I would like to produce a graph
showing the correlation between qualitative variables, quantitative
variables, and dimensions for some but not all analyses types. For
example, it would be good to see only origin, odor, and
odor.after.shaking on the graph.

By default, it doesn't seem possible to include both qualitative and
qualitative data in the correlation graph, or to filter by

I am using the code:
fviz_mfa_var(res.mfa, choice=c("quanti.var","group","quali.var"),
palette = "jco",
             col.var.sup = "violet", repel = TRUE)

I have tried using select.ind with no change. This just produces the
same graph as above:
fviz_mfa_var(res.mfa, choice=c("quanti.var","group","quali.var"),
palette = "jco",
              col.var.sup = "violet", repel = TRUE,

Attempting to simplify by just specifying one name (as a trial) does
not work either.

Is this actually possible? If so, how would it be best to attempt it?

With thanks,

[1] http://www.sthda.com/english/articles/31-principal-component-methods-in-r-practical-guide/116-mfa-multiple-factor-analysis-in-r-essentials/

Gavin Duley <gduley using gmail.com>

More information about the R-help mailing list