Hey there! Let's walk through how to solve both of your R plotting questions clearly, with practical examples using built-in datasets so you can follow along easily.

First, you'll need to calculate the mean for each group you want to plot, then pass those means to the barplot() function. Let's use the built-in mtcars dataset as an example—we'll plot average miles per gallon (mpg) grouped by number of cylinders (cyl):

# Calculate mean mpg for each cylinder group
mean_mpg_by_cyl <- tapply(mtcars$mpg, mtcars$cyl, mean)

# Create the barplot
barplot(mean_mpg_by_cyl,
        main = "Average MPG by Number of Cylinders",
        xlab = "Number of Cylinders",
        ylab = "Average MPG",
        col = "steelblue")

A quick breakdown:

• tapply() is a simple base R function to compute summary stats (here, mean) across groups. The result is a named vector where each name is a cylinder count, and the value is the corresponding mean mpg.
• We add labels, a title, and color to make the plot easier to read—feel free to tweak these to fit your data!

For ggplot2, it's best practice to first create a summary dataset that includes your group, mean value, and standard error (SE). SE is calculated as standard deviation / sqrt(sample size). We'll use dplyr (part of the tidyverse suite) to summarize the data, then plot with ggplot2:

# Load the tidyverse (includes ggplot2 and dplyr)
library(tidyverse)

# Create a summary dataset with mean and SE
summary_stats <- mtcars %>%
  group_by(cyl) %>%
  summarise(
    mean_mpg = mean(mpg),
    se_mpg = sd(mpg) / sqrt(n())  # Formula for standard error
  )

# Build the plot
ggplot(summary_stats, aes(x = factor(cyl), y = mean_mpg)) +
  geom_col(fill = "forestgreen", width = 0.7) +  # Use geom_col for precomputed means (not counts)
  geom_errorbar(aes(ymin = mean_mpg - se_mpg, ymax = mean_mpg + se_mpg),
                width = 0.2) +  # Adjust width of the error bar caps
  labs(title = "Average MPG by Cylinders (with Standard Error)",
       x = "Number of Cylinders",
       y = "Average MPG") +
  theme_minimal()

Key notes here:

• group_by() + summarise() lets us cleanly compute the mean and SE for each group. n() gives the number of observations in each group.
• geom_col() is used instead of geom_bar() because we're plotting precomputed mean values (not raw counts).
• geom_errorbar() defines the upper and lower bounds of the error bars using the SE we calculated. Adjust the width parameter to make the caps look good for your plot.
• The theme_minimal() and labs() functions polish the plot's appearance and labeling.

内容的提问来源于stack exchange，提问作者Erica Hashimoto