Add patterns to boxplots in r

Question

Good morning,

I am looking for a way to add patterns on my r boxplots. I have seen that some solutions exist already (here for exemple) but I have also find a more straitforward way to do this (here). However, I cannot manage to reproduce the exemple given hereby.

structure(list(day = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("d0", 
"db"), class = "factor"), name = structure(1:20, .Label = c("x1", 
"x2", "x3", "x4", "x5", 
"x6", "x7", "x8", "x9", 
"x10", "x11", "x12", "x13", 
"x14", "x15", "x16", "x17", 
"x18", "x19", "x20", "x21", "x22", 
"x23", "x24", "x25", "x26", 
"x27", "x28", "x29", "x30", 
"x31", "x32", "x33", "x34", 
"x35", "x36", "x37", "x38", "x39"
), class = "factor"), mean = c(36.6, 36.2636363636364, 36.9285714285714, 
35.325, 37.0692307692308, 37.2357142857143, 36.85, 36.0153846153846, 
35.6384615384615, 36.76, 36.8538461538462, 35.4307692307692, 
35.3, 36.8153846153846, 36.1666666666667, 37.775, 37.3733333333333, 
36.4666666666667, 36.6071428571429, 36.6)), row.names = c(NA, 
-20L), class = c("grouped_df", "tbl_df", "tbl", "data.frame"), vars = "day", drop = TRUE, indices = list(
    0:19), group_sizes = 20L, biggest_group_size = 20L, labels = structure(list(
    day = structure(1L, .Label = c("d0", "db"), class = "factor")), row.names = c(NA, 
-1L), class = "data.frame", vars = "day", drop = TRUE))

What I have so far are the following lines:

patternboxplot(data=temp_mean, x=temp_mean$day, y = temp_mean$mean, group = NULL,
                  pattern.type = c('nwlines', 'blank'), pattern.line.size = c(6,1), 
                  pattern.color = c('black', 'white'),
                  background.color = c('white', 'white'),
                  frame.color = c('black', 'black'),
                  density=c(6,1))

When running it, I get the following error:

"Error in imagetodf2(readPNG(paste(location, "/", pattern.type[i], ".png", : object 'xmin' not found".

If anyone has a clue about this I'll be glad to hear about it!

Thank you in advance.

EDIT: I would like to plot the following figure (obtained here using the exact same dataset and gglpot2 package)

Hi, can you elaborate on what you would like to plot? From your data frame there are no groups, and only 1 x variable (day=1).. — StupidWolf, Feb 13 '20 at 10:54
I have added the boxplot I would like to obtain but with some patterns on it. I have two groups "d0" and "db". — Blue, Feb 13 '20 at 10:59
I changed the shade by modifiying the ```background.color```, it works well. — Blue, Feb 13 '20 at 13:54

score 2 · Accepted Answer · answered Feb 13 '20 at 11:40

2

Simulate something like your data:

temp_mean = data.frame(
  day = sample(c("d0","db"),100,replace=TRUE),
  mean = rnorm(100)
)

I tried but cannot get them to have different fills, with group per x-axis break. You can try something like below (maybe not ideal):

 patternboxplot(data=temp_mean, x=1, y = temp_mean$mean, group = temp_mean$day,
               pattern.type = c('nwlines', 'blank'), pattern.line.size = c(6,1), 
               pattern.color = c('black', 'white'),
               background.color = c('white', 'white'),
               frame.color = c('black', 'black'),
               density=c(6,1),
               legend.h=30,legend.w=0.05,legend.x.pos=1.075, 
               legend.y.pos=0.88, 
               legend.pixel=10)

answered Feb 13 '20 at 11:40

StupidWolf

45,075
17
40
72

I'll have to work a bit around legend positionning but it is working! Thank you very much. Nevertheless, I wonder how putting "x=1" does the trick. Is there any explanation behind this? – Blue Feb 13 '20 at 12:48
In ggplot2, if you set for example aes(x=1,y=..,fill=group), it will create a boxplot with only 1 axis tick at 1, and two boxplots separated by groups. You are saying the x is a constant, hence this result. In patternboxplot, the group goes with the fill, so given your data, and what you need, I just tried it this way – StupidWolf Feb 13 '20 at 13:16
@Blue, how did you position the legend in the end? I tried a few times but it was not easy. You could add you final plotting command as an answer too, in case people find it usefull :) – StupidWolf Feb 13 '20 at 13:19
Thanks for the explanation (: As for the legend I'm stuck. When I change values for x and y position, only labels (texts "d0" and "db") are moving and not the "boxes" which stay at the top of the graph. – Blue Feb 13 '20 at 13:49

score 0 · Answer 2 · answered Feb 13 '20 at 09:57

0

This will not crash, however you will have to experiement yourself with the pattern of your liking.

patternboxplot(data=temp_mean, x=factor(temp_mean$day), y = as.numeric(temp_mean$mean),group=NULL,
               pattern.type = c('nwlines'), pattern.line.size = c(6,1), 
               pattern.color = c('black'),
               background.color = c('white'),
               frame.color = c('black'),
               density=c(6,1))

answered Feb 13 '20 at 09:57

Dimitrios Zacharatos

850
1
6
16

When trying the code you gave me I have the exact same error message than the one before... Could the problem comes from the pre-treatment of my data? I computed the means by using ```dplyr``` package – Blue Feb 13 '20 at 10:33

Add patterns to boxplots in r

2 Answers2