How can I separate factor levels from the variable name in modelsummary?
For instance:
library(modelsummary)
mod <- lm(mpg ~ factor(cyl), mtcars)
modelsummary(mod)
gives me:
+--------------+---------+
| | (1) |
+==============+=========+
| (Intercept) | 26.664 |
+--------------+---------+
| | (0.972) |
+--------------+---------+
| factor(cyl)6 | -6.921 |
+--------------+---------+
| | (1.558) |
+--------------+---------+
| factor(cyl)8 | -11.564 |
+--------------+---------+
| | (1.299) |
+--------------+---------+
I want:
+---------------+---------+
| | (1) |
+===============+=========+
| (Intercept) | 26.664 |
+---------------+---------+
| | (0.972) |
+---------------+---------+
| factor(cyl)_6 | -6.921 |
+---------------+---------+
| | (1.558) |
+---------------+---------+
| factor(cyl)_8 | -11.564 |
+---------------+---------+
| | (1.299) |
+---------------+---------+
EDIT: I prefer not to specify the variable name in coef_rename, but to have a solution that would work for any factor variable
The coef_rename argument of modelsummary() can take a named vector of values to rename so you can create a helper function to construct the desired names to pass to this argument. The following relies on model objects containing xlevels so won't work with models that don't.
renamer <- function(model, sep = "_") {
res <- mapply(\(x, y) setNames(paste(x, y[-1], sep = sep), paste0(x, y[-1])),
names(model$xlevels),
model$xlevels,
USE.NAMES = FALSE)
unlist(res)
}
mod <- lm(mpg ~ factor(cyl) + vs, transform(mtcars, vs = factor(vs)))
Where renamer(mod) produces:
factor(cyl)6 factor(cyl)8 vs1
"factor(cyl)_6" "factor(cyl)_8" "vs_1"
Use with modelsummary():
modelsummary(mod, coef_rename = renamer(mod), "markdown")
+---------------+---------+
| | (1) |
+===============+=========+
| (Intercept) | 27.290 |
+---------------+---------+
| | (2.086) |
+---------------+---------+
| factor(cyl)_6 | -7.153 |
+---------------+---------+
| | (1.724) |
+---------------+---------+
| factor(cyl)_8 | -12.190 |
+---------------+---------+
| | (2.262) |
+---------------+---------+
| vs_1 | -0.689 |
+---------------+---------+
| | (2.021) |
+---------------+---------+