使用ggplot2制作以零为中心的堆叠蝴蝶/龙卷风图展示Likert问卷前后数据

先处理数据，把occasion转成带标签的因子，方便区分前后：

all.gg.data <- all.gg.data %>%
  mutate(occasion = factor(occasion, levels = c(1, 2), labels = c("Before", "After")))

替换你的ggplot代码，核心调整分面方式和参数，实现水平对齐的堆叠分面：

library(ggplot2)
library(dplyr)
library(scales)

likert_viz <- ggplot(all.gg.data, aes(x = number, y = item, fill = opinion)) +
  geom_col(width = 0.8) +  # 把过宽的条形调窄，避免重叠
  geom_vline(xintercept = 50, color = "white", linewidth = 1) +  # 强化中间参考线
  scale_y_discrete(limits = unique(all.gg.data$item),  # 固定y轴问题顺序，两个分面对齐
                   labels = wrap_format(50)) +
  scale_x_continuous(labels = percent_format(scale = 1),  # 若number是小数，把scale改成100
                     breaks = seq(0, 100, 25)) +
  labs(title = "教育模块实施前后观点对比",
       subtitle = "Likert问卷结果",
       x = "百分比",
       y = "问题",
       fill = "观点") +
  facet_grid(. ~ occasion) +  # 替代facet_wrap，实现左右水平分面且y轴完全对齐
  theme_minimal() +
  theme(
    strip.text = element_text(size = 12, face = "bold"),
    plot.title = element_text(size = 14, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(hjust = 0.5)
  )

print(likert_viz)

关键修改说明： #

• 用facet_grid(. ~ occasion)替代facet_wrap：生成左右严格对齐的分面，保证同一个问题在两个分组里的y轴位置完全一致，解决分面不对齐的核心问题。
• 固定y轴顺序：用unique(all.gg.data$item)做刻度限制，避免数据重复导致的问题顺序混乱。
• 调整条形宽度：把原有的width=5.0改成0.8，让条形紧凑合理，不会互相重叠。

如果想要传统“蝴蝶图”（Before在x轴左侧负方向，After在右侧正方向，同一问题在同一y位置），可以用这个版本：

# 先调整数据：把Before的数值取反
all.gg.data <- all.gg.data %>%
  mutate(
    occasion = factor(occasion, levels = c(1, 2), labels = c("Before", "After")),
    number = ifelse(occasion == "Before", -number, number)
  )

likert_butterfly <- ggplot(all.gg.data, aes(x = number, y = item, fill = opinion)) +
  geom_col(width = 0.8) +
  geom_vline(xintercept = 0, color = "black", linewidth = 1) +
  scale_y_discrete(limits = unique(all.gg.data$item), labels = wrap_format(50)) +
  scale_x_continuous(labels = function(x) paste0(abs(x), "%"),  # 显示绝对值百分比
                     breaks = seq(-100, 100, 25)) +
  labs(title = "教育模块实施前后观点对比（蝴蝶图）",
       subtitle = "Likert问卷结果",
       x = "百分比",
       y = "问题",
       fill = "观点") +
  theme_minimal() +
  theme(
    plot.title = element_text(size = 14, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(hjust = 0.5)
  )

print(likert_butterfly)

内容的提问来源于stack exchange，提问作者Nervous Optic Head