ggplot2 custom stacked bar plot-CodePudding

I used a dataset to fit three kinds of models, and applied a decomposition to calculate feature importance of each variable. I want to visualize the results like the example below to draw a comparison across three models. Is it possible to draw a stacked bar plot with given y and decomposition results?

ggplot(data = diamonds)   
  geom_bar(mapping = aes(x = cut, fill = clarity))

Here is an example of my data:

tibble(
  methods = c("linear regression", "decision tree", "random forest"), 
  y  = c(1, 2, 3), 
  g1 = c(0.3, 0.8, 1.3), 
  g2 = c(0.5, 1,   1), 
  g3 = c(0.2, 0.2, 0.7)
)

# A tibble: 3 × 5
  methods               y    g1    g2    g3
  <chr>             <dbl> <dbl> <dbl> <dbl>
1 linear regression     1   0.3   0.5   0.2
2 decision tree         2   0.8   1     0.2
3 random forest         3   1.3   1     0.7

CodePudding user response：

You can do it with the melt function from reshape2 package to melt your tibble into a format in which each row is a unique id-variable entry:

library(reshape2);
library(ggplot2);
df_melt <- melt(df, id.vars = 'methods')
ggplot(df_melt, aes(x=methods, y=value, fill=variable))   geom_bar(position="fill", stat="identity")