factorplot (version 1.3)

fp_to_df: Convert factorplot output to data frame

Description

Converts the output from factorplot to a data frame that is convenient for custom plotting. The plot method for factorplot objects from the factorplot package will produce a plot that is lightly customisable. However, for more control over the appearance of the plot, and to plot estimates on the diagnoal of the display, returning the data and making a plot is easier.

Usage

fp_to_df(obj, type = c("upper_tri", "both_tri"), ...)

Value

A data frame with columns - `row`: The name of the parameter in the row - `col`: The name of the parameter in the column - `estimate`: The estimate for the parameter (on the diagonal) - `difference`: Pairwise differences between parameters. - `p_value`: The p-value of the pairwise difference.

Arguments

obj

An object of class `factorplot` produced by the `factorplot()` function from the `factorplot` package.

type

Indicates whether you want the resulting plot to have differences on the upper triangle and p-values on the lower triangle (if `"both_tri"`) or both p-values and differences to be plotted in the upper triangle.

...

Other arguments, currently ignored.

Examples

library(factorplot)
library(ggplot2)
data(chickwts)
mod <- lm(weight ~ feed, data=chickwts)
fp <- factorplot(mod, factor.variable = "feed", order="size")
chick_df <- fp_to_df(fp, type="upper_tri")
ggplot(chick_df, aes(x=row, y=column)) + 
  geom_tile(aes(fill=difference), color="black") + 
  geom_text(aes(label = ifelse(p_value < .05, "*", "")), color="white", size=10) + 
  scale_fill_viridis_c(option = "D", na.value = "transparent") + 
  theme_classic() + 
  geom_text(data=chick_df, 
            aes(x=row, y=column, label=round(estimate, 2))) + 
  labs(fill="Difference", x="", y="")