r - Summarising data frame with barplot of means and their standard deviation, ggplot2

Question

I have a dataset of 8 columns and 152 rows. My aim is to create a barplot using ggplot2 of each column's means and their standard deviation (these vary quite a lot). I can create a scatter plot easily, but barplot comes with multiple error messages including:

Error in barplot.default(GRPA) : 'height' must be a vector or a matrix Any suggestions or example code would be great

Example of part of data:

structure(list(ALP.B = c(80L, 37L, 52L, 36L, 39L, 48L, 71L, 81L, 
77L, 38L, 56L, 33L, 64L, 70L, 43L, 45L, 59L, 42L, 59L, 45L), 
    ALT.B = c(13L, 15L, 10L, 13L, 18L, 8L, 12L, 13L, 18L, 13L, 
    10L, 28L, 10L, 13L, 12L, 28L, 15L, 7L, 11L, 13L), AST.B = c(14L, 
    16L, 13L, 13L, 12L, 13L, 18L, 16L, 19L, 14L, 15L, 21L, 15L, 
    13L, 12L, 16L, 23L, 12L, 14L, 12L), TBL.B = c(12.654, 6.498, 
    4.788, 6.84, 14.364, 6.156, 9.063, 10.773, 7.353, 7.182, 
    7.866, 8.721, 13.338, 7.866, 11.628, 10.089, 5.301, 9.918, 
    7.353, 7.182), ALP.M = c(87L, 37L, 55L, 35L, 37L, 50L, 74L, 
    89L, 83L, 36L, 58L, 32L, 78L, 78L, 43L, 51L, 60L, 47L, 50L, 
    51L), ALT.M = c(22L, 25L, 10L, 11L, 21L, 8L, 10L, 17L, 21L, 
    16L, 13L, 27L, 14L, 18L, 13L, 41L, 14L, 8L, 13L, 14L), AST.M = c(22L, 
    23L, 13L, 12L, 15L, 13L, 15L, 13L, 22L, 17L, 18L, 27L, 16L, 
    15L, 13L, 23L, 22L, 12L, 13L, 15L), TBL.M = c(23.085, 8.037, 
    6.498, 8.037, 16.758, 5.985, 7.524, 7.866, 8.379, 7.866, 
    8.208, 13.338, 15.732, 8.208, 14.706, 15.39, 7.866, 7.353, 
    9.918, 7.866)), row.names = c(NA, 20L), class = "data.frame")

My code is rudimental, as i have tried so many:

ggplot(colMeans(GRPA), aes(x="drug", y="value")) + 
  geom_bar(stat = "identity")

Answer 1

There are several issues with your code. First ggplot2 works on data frames while you pass it a vector colMeans(GRPA). Additionally if you want pass ggplot2 the names of variables do that without quotes.

To achieve your desired result it's best to reshape your dataset into long or tidy format using e.g. tidyr::pivot_longer(). Afterwards you may use dplyr to compute the means (and/or standard deviation) per drug:

This summarised dataset can then be easily plotted via ggplot2.

library(dplyr)
library(tidyr)
library(ggplot2)

# Reshape dataset to long format, compute means per drug using group_by + summarise
GRPA_long <- GRPA %>% 
  pivot_longer(everything(), names_to = "drug", values_to = "value") %>% 
  group_by(drug) %>% 
  summarise(mean = mean(value), sd = sd(value))
#> `summarise()` ungrouping output (override with `.groups` argument)

ggplot(GRPA_long, aes(x = drug, y = mean)) + 
  geom_bar(stat = "identity") + 
  geom_errorbar(aes(ymin = mean - sd, ymax = mean + sd))