Re: [Gretl-users] Correlation heatplot
Thank, Allin! It works fine now! Of course the triangle is just an
option which I personally prefer. No need to run lower() before.
Btw, do you think this could become a public package similar to the (of
course) much more sophisticated R-package corrplot (URL:
https://cran.r-project.org/web/packages/corrplot/vignettes/corrplot-intro...
Best,
Artur
Gretl tends to assume that if we hear anything from gnuplot on
stderr,
then something has gone wrong in generating a plot. But we have a short
list of "false positives" where gnuplot squawks at nothing much. I've
added this warning to the list (in git) so the script should now run OK.
All the same, I wonder, why not just show the whole matrix rather than
just a triangle?
Allin