How to manually plot Box and Whisker graph?

2
votes

Box and Whisker graph displays the following information: max, min, mean, 75th percentile, 25th percentile. If I have these information, can I plot the corresponding B&W graph?

I have this data frame called TP.df:

        pb1    ag1     pb2     ag2     pb3    ag3
Nb      498    498      85      85      68     68
Min       0      0       0       0       0      0
Max    1.72    461   2.641   260.8     0.3    144
Mean   0.06   19.2    0.15   35.35    0.02   9.11
75_p   0.06     20    0.08      33    0.02      8
25_p   0.01     10       0      14    0.01      4

file:

,pb1,ag1,pb2,ag2,pb3,ag3
Nb,498,498,85,85,68,68
Min,0,0,0,0,0,0
Max,1.72,461,2.641,260.8,0.3,144
Mean,0.06,19.2,0.15,35.35,0.02,9.11
75_p,0.06,20,0.08,33,0.02,8
25_p,0.01,10,0,14,0.01,4

How can I have the corresponding Box and Whisker graph:

  • it must show 6 boxes
  • x axis : pb1, ag1, pb2, ag2, pb3, ag3
  • y axis : 0 to max(TP.df[Max,])
3
rplotboxplot
Did you look at ? boxplot and ? boxplot.stats? - Thomas

3 Answers

7
votes

Use bxp to make a boxplot from a five-number summary.

Does this do the sort of thing you want?

testdata=data.frame(R1=c(0,5,3,2,4),R2=c(1,7,3,2.8,6))
o=c(1,5,3,4,2) # the rows in increasing order
bxp.data=list(stats=data.matrix(testdata[o,]),n=rep(1,ncol(testdata)))
# the n=... parameter doesn't affect the plot, but it still needs to be there
bxp(bxp.data)

You can use standard graphical parameters (see help(par)) to label the axes and make it look pretty.

(Note: row numbers are off by one because the question was edited after I posted this answer. I'll leave the answer as is so that it matches the first two comments.)

1
votes

that might be a solution:

  #create a dummy boxplot that you can modify the data easily
  z<- boxplot(1:10)

  #look at the outbut an assign yout data to stats
  z$stats<- your_data

  #use bxp to plot, via add you can combine all three
  bxp(z)

or use the bxp function from the beginning.. see also: http://r.789695.n4.nabble.com/Box-plot-with-5th-and-95th-percentiles-instead-of-1-5-IQR-problems-implementing-an-existing-solution-td3456123.html

ben

0
votes

Using manual boxplot with ggplot, you can manipulate plot aesthetics and add other parameters:

ggplot()+geom_boxplot(aes(x=1, y = 2:5, lower = 3, upper = 4, middle = 3.5, ymin=2, ymax=5))