Is it possible to plot a boxplot from previously-calculated statistics easily (in R?) [duplicate]

9
votes

Possible Duplicate:
Draw bloxplots in R given 25,50,75 percentiles and min and max values

I have a number of sets of summary statistics from various datasets: max, min, mean, median etc. I'd like to plot box-plots of these - or at least, similar plots to boxplots (I don't have UQ and LQ stats, although it may be possible to get those).

I don't have the original data, so I can't just use the boxplot function in R. Is there an easy way to do this in R when you just have the summary statistics? If not, is there an easy way to do this using another free tool?

1
rdata-visualizationboxplot
As this is a R-specific programming question, stackoverflow may have been the more suitable site (there's an [r] tag at SO, too). - cbeleites unhappy with SX
It doesn't read like an R-specific question to me: it seeks a method to draw boxplots from summary statistics. - whuber
This is a duplicate. I like this question, but it was closed as a duplicate of this question - GSee

1 Answers

20
votes

The boxplot function in R uses a low-level function called bxp which accepts summary statistics. A simple example (lower whisker=1, 1st quartile=2, median=3, 3rd quartile=4, upper whisker=5) would look like this:

summarydata<-list(stats=matrix(c(1,2,3,4,5),5,1), n=10)
bxp(summarydata)

If you want to know more about the data structure that bxp accepts as input, look at the return value of the high-level boxplot function for some dummy data, i.e. try

sd<-boxplot(dummydata)
str(sd)