r - Floating barcharts

Question

I want to make bar charts where the bar minimum can be specified (much like the box in a box and whisker plot). Can barplot do that? I suspect the answer's in ggplot, but I can't find an example. Here's some data:

                    X  Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Oct  Nov  Dec
1      Highest recorded 31.5 31.8 30.3 28.0 24.9 24.4 21.7 20.9 24.5 25.4 26.0 28.7
2  Mean monthly maximum 27.8 28.6 27.0 24.8 22.0 20.0 18.9 18.8 20.4 22.4 23.9 26.8
3    Mean daily maximum 24.2 24.8 23.1 20.9 18.4 16.3 15.5 15.7 16.9 18.3 20.0 22.4
4                  Mean 19.1 19.8 18.1 16.2 13.8 11.9 11.2 11.6 12.7 14.1 15.7 17.7
5    Mean daily minimum 14.0 14.7 13.1 11.4  9.2  7.5  6.9  7.4  8.4 10.0 11.4 13.0
6  Mean monthly minimum  7.6  9.1  6.8  3.8  2.3 -0.5 -0.2  1.0  2.3  3.7  5.3  6.7
7       Lowest recorded  4.0  5.6  4.1 -1.3  0.0 -3.1 -2.6 -1.4 -0.8  2.0  2.7  4.1

xaxis =c("J" ,"F" ,"M" ,"A" ,"M" ,"J","J","A", "S", "O","N","D")

So ideally, I end up with a stacked bar for each month, that starts at the 'Lowest recorded' value, rather than at zero.

I've also had a try with superbarplot from the UsingR package. I can get the bars to start where I want, but can't move the x axis down out of the centre of the plot. Thanks in advance.

Answer 1

You can use geom_boxplot in ggplot2 to get what (I think) you want specifying the precomputed values and stat = 'identity' and use geom_crossbar to put in the other

# first, your data
weather <- read.table(text = 'X  Jan  Feb  Mar  Apr  May  Jun  Jul  Aug  Sep  Oct  Nov  Dec
1      "Highest recorded" 31.5 31.8 30.3 28.0 24.9 24.4 21.7 20.9 24.5 25.4 26.0 28.7
2  "Mean monthly maximum" 27.8 28.6 27.0 24.8 22.0 20.0 18.9 18.8 20.4 22.4 23.9 26.8
3    "Mean daily maximum" 24.2 24.8 23.1 20.9 18.4 16.3 15.5 15.7 16.9 18.3 20.0 22.4
4                  "Mean" 19.1 19.8 18.1 16.2 13.8 11.9 11.2 11.6 12.7 14.1 15.7 17.7
5   "Mean daily minimum" 14.0 14.7 13.1 11.4  9.2  7.5  6.9  7.4  8.4 10.0 11.4 13.0
6  "Mean monthly minimum"  7.6  9.1  6.8  3.8  2.3 -0.5 -0.2  1.0  2.3  3.7  5.3  6.7
7       "Lowest recorded"  4.0  5.6  4.1 -1.3  0.0 -3.1 -2.6 -1.4 -0.8  2.0  2.7  4.1', header =T)


library(reshape2)
library(ggplot2)
# reshape to wide format (basically transposing the data.frame)
w <- dcast(melt(weather), variable~X)
ggplot(w, aes(x=variable,ymin = `Lowest recorded`, 
         ymax = `Highest recorded`, lower = `Lowest recorded`, 
         upper = `Highest recorded`, middle = `Mean daily maximum`)) + 
     geom_boxplot(stat = 'identity') +
     xlab('month') + 
     ylab('Temperature') +
     geom_crossbar(aes(y = `Mean monthly maximum` ))+
     geom_crossbar(aes(y = `Mean monthly minimum`)) +
     geom_crossbar(aes(y = `Mean daily maximum` ))+
     geom_crossbar(aes(y = `Mean daily minimum`)) 

This is partially described in an example in the help for geom_boxplot