Create boxplots and barplots from dataframe with +2 columns

General
, ,
1

Hello again. I am trying to create boxplots and barplots from different sets of data I have defined. Problem is, in all guides I have found on how to do it, only two variables are used, but in my data frames I have either 3 or 5. Here are a couple examples of said dataframes:

summary(ChloroCont)
Art. Mean_chloro_May Mean_chloro_June Mean_chloro_July Mean_chloro_August Mean_chloro_September
Length:18 Min. :11.68 Min. :19.22 Min. :24.54 Min. :24.44 Min. :23.47
Class :character 1st Qu.:17.86 1st Qu.:26.26 1st Qu.:28.74 1st Qu.:29.34 1st Qu.:29.18
Mode :character Median :19.50 Median :30.30 Median :33.56 Median :36.72 Median :32.70
Mean :21.19 Mean :29.46 Mean :33.16 Mean :34.42 Mean :34.10
3rd Qu.:25.46 3rd Qu.:33.04 3rd Qu.:37.56 3rd Qu.:38.64 3rd Qu.:37.64
Max. :31.56 Max. :36.47 Max. :42.73 Max. :42.88 Max. :46.97
NA's :3
summary(MeanDBH)
Art. Mean_DBH_May Mean_DBH_September
Length:18 Min. : 8.52 Min. : 8.76
Class :character 1st Qu.:10.09 1st Qu.:10.99
Mode :character Median :11.04 Median :11.83
Mean :12.12 Mean :12.76
3rd Qu.:13.91 3rd Qu.:14.14
Max. :21.64 Max. :20.13

I am quite frustrated because this is an incredibly easy task to perform in Excel, but I am required to do this in R by my supervisor. Is there any way to create boxplots/barplots with dataframes like this that doesn't involve modifying the original dataframes or creating new ones?

2

Hello,

I'm not sure I fully understand what you mean, but if you like to plot multiple variables to different boxplots you can do that like this:

set.seed(4) #For reproducibility

### BASE R ###

#Generate random data
myData = data.frame(x = LETTERS[1:3], y = runif(300))
head(myData)
#>   x           y
#> 1 A 0.585800305
#> 2 B 0.008945796
#> 3 C 0.293739612
#> 4 A 0.277374958
#> 5 B 0.813574215
#> 6 C 0.260427771

#Boxplot
boxplot(y ~ x, data = myData)

672×480 2.4 KB

### TIDYVERSE ###

library(tidyverse)

#Generate random data
myData = data.frame(
  x = runif(20, 0, 1), 
  y = runif(20, 1, 2),
  z = runif(20, -1, 1)
)
head(myData)
#>           x        y          z
#> 1 0.8097525 1.633837  0.9275156
#> 2 0.5377679 1.177658 -0.9663170
#> 3 0.2071941 1.688226 -0.3674301
#> 4 0.1999695 1.333764 -0.7346018
#> 5 0.9381336 1.175669 -0.4479980
#> 6 0.9736611 1.884520 -0.5123024

#Convert it to a long format
myData = myData %>% pivot_longer(everything())
head(myData)
#> # A tibble: 6 × 2
#>   name   value
#>   <chr>  <dbl>
#> 1 x      0.810
#> 2 y      1.63 
#> 3 z      0.928
#> 4 x      0.538
#> 5 y      1.18 
#> 6 z     -0.966

#Plot using ggplot
ggplot(myData, aes(y= value, group = name)) + 
  geom_boxplot()

672×480 2.14 KB

Created on 2023-01-12 by the reprex package (v2.0.1)

Hope this helps,
PJ

3

Thank you for your quick answer. I am aware you can create boxplots and barplots like these, but I was wondering if there was any way to add an additional layer to it. I am working whit monthly measurements per species, so I wanted to creat a plot that could display all monthly measurement per species without having to create one plot per month.

4

PJ seems to have given you that ; what is missing ?
here is an extension with a second dataframe integrated.
this is simply repeating the data prep, and vertically stacking the data before charting.

library(tidyverse)

library(tidyverse)

#Generate random data
(ChloroCont = data.frame(
  mean_chloro_may = runif(20, 0, 1), 
  mean_chloro_june = runif(20, 1, 2),
  mean_chloro_july = runif(20, -1, 1)
))

(MeanDBH = data.frame(
  Mean_DBH_May = rnorm(20),
  Mean_DBH_September=rnorm(20,mean=1)
))

#Convert them to a long format
(long_1 = ChloroCont %>% pivot_longer(everything()))
(long_2 = MeanDBH %>% pivot_longer(everything()))

stacked <- bind_rows(long_1,long_2)
#Plot using ggplot
ggplot(stacked, aes(x=name,y= value, group = name)) + 
  geom_boxplot()

image
image1022×588 28.9 KB

5

these 3 steps, and variations of these you might come up with can also be more elegantly reduced to

stacked <- map_dfr(list(
  ChloroCont,
  MeanDBH
), \(x)pivot_longer(x, everything()))
6

Something like this?

myData = data.frame(
  x = runif(20, 0, 1), 
  y = runif(20, 1, 2),
  z = runif(20, -1, 1),
  month = sample(month.abb[1:12], 20, replace = TRUE)
)

myData  <-  myData %>% pivot_longer(!month)
ggplot(myData, aes(name, value, colour = as.factor(month))) + geom_boxplot()
ggplot(myData, aes(name, value, colour = month)) + geom_boxplot(show.legend = FALSE ) +
      facet_wrap(vars(month) )
7

I am sorry, I did not see it in the previous reply. I see it now in yours. I will try to get it to work on my own based on your code. Thank you very much for your help.