On Sat, Jul 25, 2020 at 12:40 AM Martin Maechler
<maech...@stat.math.ethz.ch> wrote:
> Good answers to this question will depend very much on how many
> 'Machine' and 'Region' levels there are.
I second that.
And unless I missed something, the OP hasn't answered this question, as such.
But "10k+" combinations, does imply around 100 levels each.
Another important question is, are the combinations unique or not?
It would be possible to create an (approx):
100x100 heatmap of boolean values, for unique combinations, or;
100x100 heatmap of counts (or density), for non-unique combinations.
But unless there's some meaningful order to the levels, the resulting
plot may end up looking like a $3 pizza.
I'm unable to comment on possible exploratory value, but I doubt that
this is a good approach, for presentation purposes.
If the goal was some sort of ranking, a textual summary, may work better...?
Or you could plot relevant subsets of the data...
______________________________________________
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
