How can I add significance to the correlation without overlap between the symbol (*) and the correlation value ??
For this, I use the PerformanceAnalytics
package. See:
PerformanceAnalytics::chart.Correlation(iris[, -5])
Viacorrplot
isnotpossible/wouldhavetodoalotofprogramming. In this question , they present an option via corrgram
but it is not practical in my opinion. If it were to use corrplot
I would use something like this:
m.cor <- cor(iris[, -5])
m.sig <- cor.mtest(iris[, -5])$p
corrplot(corr = m.cor, outline = F, p.mat = m.sig, tl.srt = 0,
addCoef.col = "black", method = "square", type = "upper", diag = F, insig = "pch")
Wherethe"X" represents that the coefficient was not nonzero (not significant).