Changing the number of significant digits shown with stat_cor

[raedevan6]

Hello, forgive me if this an easy fix. I'm trying to change the number of digits shown on the plot for both R2 and pvals. Below is my code, I would like to only show 3 sig digits.

ggplot(pxsam, aes(x=cbi, y=value, color=phy)) + geom_point() +
geom_smooth(method = "lm") + theme_pubr() +
ggpubr::stat_cor(
aes(label = paste(..rr.label.., ..p.label.., sep = ",")))

[icer182]

Hi @raedevan6. I figured this out just by using stat_poly_eq. However, I am not sure how to independently adjust the R^2 values now. I know I can adjust them together, but I'd like for them to event out by declaring rr.digits 2 for one regression line and rr.digits = 3 for the other regression line. @kassambara

[kassambara]

Related issue: #186

[kassambara]

New arguments digits added.

Applying different digits for each regression line on the same plot is not supported by stat_cor.

I don't know how to achieve this automatically in ggplot context. I would suggest the following (manual) solution:

library(ggpubr)
library(dplyr)

df <- mtcars
paste("italic(R)^2", ..rr.., sep = "~`=`~")
#> Error in paste("italic(R)^2", ..rr.., sep = "~`=`~"): objet '..rr..' introuvable
df$cyl <- as.factor(df$cyl)

# Prepare data for each group
grp1.df <- df %>% filter(cyl == "4")
grp2.df <- df %>% filter(cyl == "6")
grp3.df <- df %>% filter(cyl == "8")

# Plot
ggplot(df, aes(wt, mpg, color = cyl)) +
  geom_point() +
  geom_smooth(method = "lm") +
  stat_cor(label.x = 3, label.y = 34, digits = 3, data = grp1.df)+
  stat_cor(label.x = 3, label.y = 32, digits = 2, data = grp2.df) +
  stat_cor(label.x = 3, label.y = 30, digits = 1, data = grp3.df)

^{Created on 2020-02-12 by the reprex package (v0.3.0)}

[icer182]

Hi @kassambara. Thanks for implementing the digits variable to stat_cor. I was just curious if there was a method to keep trailing 0s? I specified digits = 2, but it deletes my trailing 0. Thank you.

[kassambara]

Hi,

for advanced p-value formatting, I would suggest the function scales::pvalue() using the option accuracy. Use accuracy = 0.01 to show 2 decimal places of precision.

library(ggpubr)
# Custom formatting fucntion
format_pval <- function(pval){
  pval <- scales::pvalue(pval, accuracy= 0.001, add_p = TRUE)
  gsub(pattern = "(=|<)", replacement = " \\1 ", x = pval)
}
# Display formatted p-value
ggscatter(mtcars, x = "wt", y = "mpg", add = "reg.line") +
  stat_cor(
    aes(label = paste(..rr.label.., format_pval(..p..), sep = "*`,`~")),
    label.x = 3
  )
#> `geom_smooth()` using formula 'y ~ x'

^{Created on 2020-05-09 by the reprex package (v0.3.0.9001)}

[icer182]

Hi @kassambara. Thanks for the explanation. However, I am looking to keep a trailing 0 in the R^2 value. Is there a current method of doing this? Sorry for the confusion.

[kassambara]

With the latest dev version, you can now specify the argument p.accuracy and r.accuracy:

suppressPackageStartupMessages(library(ggpubr))
# Default plot
sp <- ggscatter(mtcars, x = "wt", y = "mpg", add = "reg.line")
sp + stat_cor(label.x = 3)
#> `geom_smooth()` using formula 'y ~ x'

# Specify the number of decimal places of precision for 
# the p-value and the correlation coefficient r
sp + stat_cor(label.x = 3, p.accuracy = 0.001, r.accuracy = 0.01)
#> `geom_smooth()` using formula 'y ~ x'

^{Created on 2020-05-10 by the reprex package (v0.3.0.9001)}

[courtney-bryce-hilton]

Hi @kassambara. The above suggestion to including trailing zeros for the correlation coefficient by setting r.accuracy isn't working for me. I have tried both the current CRAN version of ggpubr and the dev version.

For my case, the R2 always shows as 0.6 in the plot rather than 0.60. From what I can gather, ..rr.label.. is correctly set at "0.60" but when it gets passed into the R expression format, that removes the trailing zero. Do you see any way around that?

Here's a reproducible example that manually sets a number in the label to be 0.60 to demonstrate:

sp <- ggscatter(mtcars, x = "wt", y = "mpg", add = "reg.line")
sp + stat_cor(aes(label = paste(0.60, ..p.label.., sep = "~`,`~")), label.x = 3, r.accuracy = 0.01)

Thanks