bar chart in ggplot-CodePudding

I have the following data: enter image description here

I want to have a bar chart where for each group I have two bars (IT and PT) side by side: this is my script:

 ggplot(score, aes(group, IT, fill = stain))  
    geom_bar(stat="identity", position = "fill")   
      scale_y_continuous(labels = scales::percent_format())  
    ggplot(score, aes(group, PT, fill = stain))  
     geom_bar(stat="identity", position = "fill") 
    scale_fill_manual(values = c("#edae49", "#d1495b", "#00798c", "#30638e")) 
      scale_y_continuous(labels = scales::percent_format())   
     ggtitle("CD4")

how can I put the PT and IT on the same y-axis. Also how can I keep 0% values on the plot without having the auto NA instead?

CodePudding user response：

It will make life easier if you start by pivoting your data into long format. That is, instead of having one column for IT and one for PT, you have a single column with all the values, and another column labeling the value in that row as coming from either PT or IT. You can do this easily with pivot_longer.

It sounds like you want to stack the bars according to group, but have them dodged according to IT / PT. You can't both stack and dodge directly, but you can fake it with facets:

Your code might then look something like this:

library(tidyverse)

score %>%
  pivot_longer(IT:PT, names_to = 'IT_PT') %>%
  mutate(stain = factor(stain, unique(stain)),
         group = factor(group, unique(group))) %>%
  ggplot(aes(IT_PT, value, fill = group))  
  geom_col(position = 'fill')  
  theme_minimal(base_size = 16)  
  facet_grid(.~stain, switch = 'x')  
  scale_fill_manual(name = NULL, values =  c("#edae49", "#d1495b", 
                                             "#00798c", "#30638e"))  
  scale_x_discrete(labels = c('', ''), name = 'Stain', 
                   expand = c(0.1, 0.7))  
  scale_y_continuous(labels = scales::percent_format())   
  theme(panel.spacing.x = unit(0, 'mm'))

Data used

Obviously, I don't have your data, so I had to transcribe it manually from the picture in your question.

score <- structure(list(group = c("All_BrM", "All_BrM", "All_BrM", "All_BrM", 
"Sync", "Sync", "Sync", "Sync", "Meta", "Meta", "Meta", "Meta", 
"Poly", "Poly", "Poly", "Poly"), stain = c("0%", "<10%", "11-40%", 
">40%", "0%", "<10%", "11-40%", ">40%", "0%", "<10%", "11-40%", 
">40%", "0%", "<10%", "11-40%", ">40%"), IT = c(20, 35, 1, 0, 
2, 13, 1, 0, 9, 9, 0, 0, 9, 13, 0, 0), PT = c(9, 30, 6, 0, 1, 
9, 1, 0, 4, 10, 2, 0, 4, 11, 3, 0)), class = "data.frame", 
row.names = c(NA, -16L))

CodePudding user response：

@Allan Cameron

score$group <- factor(score$group, levels=c("All_BrM", "Sync","Meta","Poly" ))
  score$stain <- factor(score$stain, levels=c("none", "<10%","10-40%",">40%" ))
  
  ggplot(score, aes(group, IT, fill = stain))  
    geom_bar(stat="identity", position = "fill")   
      scale_fill_manual(values = c("#ded9e2", "#c0b9dd", "#80a1d4", "#75c9c8")) 
      theme_prism(base_size = 9) 
      scale_y_continuous(labels = scales::percent_format())  
      theme(legend.position = "none") 
      theme(axis.text.x  = element_text(size = 10, angle=50)) 
    ggplot(score, aes(group, PT, fill = stain))  
     geom_bar(stat="identity", position = "fill") 
   scale_fill_manual(values = c("#ded9e2", "#c0b9dd", "#80a1d4", "#75c9c8")) 
      scale_y_continuous(labels = scales::percent_format())               
      theme_prism(base_size = 9) 
     theme(axis.text.x  = element_text(size = 10, angle=50))