Create several exposure combined density/ distribution plot in R-CodePudding

set.seed(42)
n=1000
db = data.frame(id=1:n,
                exp_1 = as.numeric(rnorm(n)),
                exp_2 = as.numeric(rnorm(n)),
                exp_3 = as.numeric(rnorm(n)), 
                exp_4=as.numeric(rnorm(n)))
label(db$exp_1)="Myx"
label(db$exp_2)="ff3"
label(db$exp_3)="poison-untitled"
label(db$exp_4)="NH3"

I want to create a combined density plot - with legends referring to the variable label a plot similar to

CodePudding user response：

First you could change your data to a longer format using pivot_longer from tidyr. The graph you mentioned looks like ggpubr so you can use ggdensity with scale_fill_discrete to modify the legend like this:

set.seed(42)
n=1000
db = data.frame(id=1:n,
                exp_1 = as.numeric(rnorm(n)),
                exp_2 = as.numeric(rnorm(n)),
                exp_3 = as.numeric(rnorm(n)), 
                exp_4=as.numeric(rnorm(n)))

library(ggpubr)
library(tidyr)
library(dplyr)
db %>%
  pivot_longer(cols = -id) %>%
  ggdensity(x = 'value', fill = 'name')  
  scale_fill_discrete('', labels = c('Myx', 'ff3', 'poison-untitled', 'NH3'))

^{Created on 2022-12-03 with reprex v2.0.2}

If you want to change the transparency, you could use the argument alpha in ggdensity. For more info check this documentation.