Changing color of density plots in ggplot2 based on x axis

Question

I have some data I would like to visualize and I am able to make the chart on the left, but I would like to be able to provide more information to the viewer by implementing features like those in the image to the right: shaded based on predefined ranges, and percentage of area in each range.

I recognize that this question is similar to these two answers, however I don't understand densities enough to get the dataframe in the correct format:

• ggplot2 shade area under density curve by group

• Shade (fill or color) area under density curve by quantile

Here is the code that replicates my example.

If you can, please use dplyr in your response.

Thank you in advance.

library(dplyr)
library(ggplot2)
options(scipen = 999)

#Get percentages
  diamonds%>%
    mutate(Cut = cut,
           Prices = cut(price, 
                        breaks=c(0,2499,4999, 19000), 
                        include.lowest=T, dig.lab=10))%>%
    group_by(Cut, Prices)%>%
      summarise(Count = n())%>%
    group_by(Cut)%>%
      mutate(Pct = round(Count/sum(Count), 2))%>%
    ungroup()


#Plot
  ggplot(diamonds, aes(x=price))+
    geom_density(fill="grey50")+
    facet_grid(cut~.)+
    geom_vline(xintercept = c(2500,5000))+
    theme(axis.text.y = element_blank(),
          axis.ticks.y = element_blank())

Answer 1

The problem is that you don't have density data in the diamonds data.frame. The other problem that you face is that you need to keep the facet information. I'm not sure how to get dplyr to group by cut and get density(). One can generate summary data as you did but in order to make a density plot you're going to need x,y information for each point.

One workaround I found is making the density plot p like this

p<-ggplot(diamonds,aes(x=price))+geom_density()+facet_wrap(~cut,nrow=5)

And then ussing ggplot_build function to get the data that is being plotted

pg <- ggplot_build(p)

This gets you a list where the first element is the actual dataset

pg_data<-data.frame(pg$data[[1]],stringsAsFactors = F)

You can check that you're interested in y column (which is the same as density) x is going to be the price and PANEL is going to be the facet. I didn't change this to the factor mode with Good, Very Good... but I'm guessing you can.

head(pg_data)
              y        x       density    scaled      count    n PANEL group
1 0.00005370272 326.0000 0.00005370272 0.2756038 0.08646139 1610     1    -1
2 0.00005829975 362.1977 0.00005829975 0.2991959 0.09386259 1610     1    -1
3 0.00006307436 398.3953 0.00006307436 0.3236993 0.10154972 1610     1    -1
4 0.00006798165 434.5930 0.00006798165 0.3488836 0.10945045 1610     1    -1
5 0.00007298816 470.7906 0.00007298816 0.3745772 0.11751094 1610     1    -1
6 0.00007807049 506.9883 0.00007807049 0.4006598 0.12569348 1610     1    -1
  ymin          ymax fill weight colour alpha size linetype
1    0 0.00005370272   NA      1  black    NA  0.5        1
2    0 0.00005829975   NA      1  black    NA  0.5        1
3    0 0.00006307436   NA      1  black    NA  0.5        1
4    0 0.00006798165   NA      1  black    NA  0.5        1
5    0 0.00007298816   NA      1  black    NA  0.5        1
6    0 0.00007807049   NA      1  black    NA  0.5        1

Now we can plot again everything but using the density data we needed

ggplot(data=pg_data,aes(x,y))+geom_line()+facet_wrap(~PANEL,nrow=5)+geom_area(data=subset(pg_data,x>2499&x<5000),aes(x,y),fill = "red", alpha = 0.5)