Trying to apply color gradient on histogram in ggplot

Question

I am floundering with colors in ggplot. I am trying to apply a color gradient based on the rank column below. I am pretty sure this is a discrepancy between color and fill or discrete and continuous variables. I want colors as shown on the scale in "c" and "d" below, but my closest attempts are "e" and "f" where the points are colored but not colored by gradient. The gradient I prefer applies to values of rank 1:100, with all others values' points black.

Any assistance would be most appreciated.

library(reshape2)
library(ggplot2)

co2 <- read.table(
  header=TRUE, text='
rank tons
1     2 1.00
2     4 1.00
3     7 0.00
4    44 0.00
5   104 0.00
6    48 0.05
7    32 0.50
8     5 0.00
9    78 1.00
10   12 0.00
11   15 0.00
12  176 1.00
13  440 0.02
14  249 0.00
15  481 0.00
16  388 0.00
17  458 0.05
18  488 0.00
19  264 0.00
20  203 0.00
            ')      

I tried:

#does not add rank as a color
c<- ggplot(data=co2, aes(x = tons, color=rank)) 
c + geom_dotplot(stackgroups = TRUE, binwidth = .05, binpositions = "all")  +
    scale_colour_gradient(limits=c(1, 500))

#also does not add rank as color
d<- ggplot(data=co2, aes(x = tons, color=rank)) 
d + geom_dotplot(stackgroups = TRUE, binwidth = 0.05, method = "histodot")  + 
    scale_colour_gradient(limits=c(1, 100))

#create breaks for fill-- works correctly but no gradient
co2$brks<- cut(co2$rank, c(seq(0, 100, 20), max(co2$rank)))
e<- ggplot(data=co2, aes(x = tons, fill=brks)) 
e + geom_dotplot(stackgroups = TRUE, binwidth = 0.05, method = "histodot")  

#also works correctly but no gradient
f<- ggplot(data=co2, aes(x = tons, fill=brks)) + geom_histogram() 
f 

I checked these already but I'm still missing something:

• Color Gradients With ggplot
• Use histogram breaks to apply function over second column
• gradient breaks in a ggplot stat_bin2d plot
• http://docs.ggplot2.org/current/geom_dotplot.html

Answer 1

This is a bit of a hacky answer, but it works:

##Define breaks
co2$brks<- cut(co2$rank, c(seq(0, 100, 5), max(co2$rank)))
#Create a plot object:
g = ggplot(data=co2, aes(x = tons, fill=brks)) +
   geom_dotplot(stackgroups = TRUE, binwidth = 0.05, method = "histodot")

Now we manually specify the colours to use as a palette:

 g + scale_fill_manual(values=colorRampPalette(c("white", "red"))( length(co2$brks) ))

Answer 2

I just needed to add

+ scale_fill_brewer(palette="RdYlBu")

See below:

library(reshape2)
library(ggplot2)

co2 <- read.table(
  header=TRUE, text='
rank tons
1     2 1.00
2     4 1.00
3     7 0.00
4    44 0.00
5   104 0.00
6    48 0.05
7    32 0.50
8     5 0.00
9    78 1.00
10   12 0.00
11   15 0.00
12  176 1.00
13  440 0.02
14  249 0.00
15  481 0.00
16  388 0.00
17  458 0.05
18  488 0.00
19  264 0.00
20  203 0.00
            ')                                                                                                          

#create breaks for fill--
co2$brks<- cut(co2$rank, c(seq(0, 100, 20), max(co2$rank)))
e<- ggplot(data=co2, aes(x = tons, fill=brks)) 
e + geom_dotplot(stackgroups = TRUE, binwidth = 0.05, method = "histodot")  + scale_fill_brewer(palette="RdYlBu")

#also works correctly!
ggplot(data=co2, aes(x = tons, fill=brks)) + geom_histogram() + scale_fill_brewer(palette="RdYlBu")