求助：柱状图中标准差误差线未居中的修复及替代方案咨询

使用R绘制带均值和标准差的柱状图时，标准差误差线未与柱状条居中对齐。示例代码如下：

# Example data: Values and their standard deviations
values <- c(10, 15, 20, 25, 30)  # Mean values
std_dev <- c(1, 2, 1.5, 3, 2.5)   # Standard deviations
labels <- c("A", "B", "C", "D", "E")  # Labels for each bar

# Create a bar plot with custom labels and error bars
barplot(values, ylim = c(0, max(values) + max(std_dev)), 
        names.arg = labels, xlab = "Groups", ylab = "Values", 
        main = "Example Bar Plot with Error Bars", cex.names = 0.7,
        col = "skyblue", cex.axis = 1)

# Add error bars representing standard deviations
for (i in 1:length(values)) {
  arrows(x0 = i, y0 = values[i] - std_dev[i], 
         x1 = i, y1 = values[i] + std_dev[i], 
         angle = 90, code = 3, length = 0.1)
}

问题原因 #

barplot()函数返回的是每个柱子中心的实际x坐标，并非简单的1、2、3...序列。直接用循环变量i作为误差线的x位置，会导致误差线和柱子错位。

修复方法1：捕获barplot的x坐标 #

修改代码，保存barplot()返回的坐标，用这个坐标绘制误差线：

# Example data: Values and their standard deviations
values <- c(10, 15, 20, 25, 30)  # Mean values
std_dev <- c(1, 2, 1.5, 3, 2.5)   # Standard deviations
labels <- c("A", "B", "C", "D", "E")  # Labels for each bar

# 捕获每个柱子的中心x坐标
bar_pos <- barplot(values, ylim = c(0, max(values) + max(std_dev)), 
        names.arg = labels, xlab = "Groups", ylab = "Values", 
        main = "Example Bar Plot with Error Bars", cex.names = 0.7,
        col = "skyblue", cex.axis = 1)

# 使用捕获的坐标绘制居中的误差线
for (i in 1:length(values)) {
  arrows(x0 = bar_pos[i], y0 = values[i] - std_dev[i], 
         x1 = bar_pos[i], y1 = values[i] + std_dev[i], 
         angle = 90, code = 3, length = 0.1)
}

替代方法：用ggplot2绘制（更简洁） #

ggplot2会自动处理误差线与柱子的对齐，代码更直观易维护：

library(ggplot2)

# 整理数据为数据框格式
df <- data.frame(
  group = labels,
  mean = values,
  sd = std_dev
)

# 绘制带误差线的柱状图
ggplot(df, aes(x = group, y = mean)) +
  geom_col(fill = "skyblue") +
  geom_errorbar(aes(ymin = mean - sd, ymax = mean + sd), 
                width = 0.2) +
  labs(x = "Groups", y = "Values", 
       title = "Example Bar Plot with Error Bars") +
  theme_bw()

内容的提问来源于stack exchange，提问作者Reda