这取决于你的数据是如何格式化的。我在下面提供了两个例子。
如果从包含汇总值的表开始，则可以执行此操作

library(ggplot2)

levels <- c("veryactive", "fairlyactive", "lightlyactive", "sedentary")
df1 <- data.frame(activitylevel = factor(levels, levels = levels),
                 meancalories = c(3000, 2500, 2000, 1500))
ggplot(df1, aes(x = activitylevel, y = meancalories)) +
  geom_col()

由reprex package（v2.0.1）于2023年1月30日创建
如果你从原始数据开始，你可以这么做。

library(ggplot2)
levels <- c("veryactive", "fairlyactive", "lightlyactive", "sedentary")
df2 <- data.frame(activitylevel = factor(rep(levels,
                                     each = 20), levels = levels),
                 calories = c(rnorm(20, 3000, 100),
                              rnorm(20, 2500, 100),
                              rnorm(20, 2000, 100),
                              rnorm(20, 1500, 100))
                 )
ggplot(df2, aes(x = activitylevel, y = calories)) +
  stat_summary(geom = "col", fun = "mean")

由reprex package（v2.0.1）于2023年1月30日创建
最后，如果您开始使用宽格式的数据（即每个活动级别对应一列），那么我建议您查找函数tidyr::pivot_longer，它将把您的数据转换为stat_summary所需的格式。

使用colMeans

cols <- c("veryactive", "fairlyactive", "lightlyactive", "sedentary")

# base R
barplot(colMeans(Activity_Zero[, cols]))

# ggplot
library(ggplot2)

ggplot(stack(colMeans(Activity_Zero[, cols])), aes(ind, values)) + geom_col()

您可能需要像这样透视数据：

library(tidyverse)

df <- tibble(
  sedentary = sample(10:50, 100, replace = TRUE),
  lightlyactive = sample(10:20, 100, replace = TRUE),
  fairlyactive = sample(100:400, 100, replace = TRUE),
  veryactive = sample(300:600, 100, replace = TRUE),
  calories = sample(1500:2500, 100, replace = TRUE)
)

df |> 
  pivot_longer(c(veryactive, fairlyactive, lightlyactive, sedentary), 
               names_to = "group",
               values_to = "vals") |> 
  ggplot(aes(fct_inorder(group), vals)) +
  stat_summary(geom = "bar", fun = mean)