如何在R中按年份差值重新排序点图（Dot Plot）

数据集 #

我有1992至1999年间5个城市的居民汽车拥有率百分比数据，数据集的R代码如下：

city <- c("A", "A", "B", "B", "F", "F", "T", "T", 
          "C", "C") 
year <- c("1992", "1999", "1992", "1999", "1992", "1999", "1992", "1999", "1992", "1999")
# 汽车拥有率
car_owner <- c("15.9", "75.6", "35.8", "95.2", "15.9", "74", "6", "64", "41", "97.9")

df <- data.frame(city, year, car_owner)
df$car_owner <- as.numeric(df$car_owner)

现有绘图代码 #

我用以下代码绘制了点图，已展示全部信息，但需要调整城市排序：

df %>%
  group_by(city) %>%
  mutate(Difference = car_owner - lead(car_owner),
         Position = car_owner - (0.5 * Difference)) %>%
  ggplot(aes(x = car_owner, y = city))+
  geom_vline(xintercept = c(),
             lty = 2, alpha = 0.5) +
  geom_line(aes(group = city)) +
  geom_point(aes(color = year), size=3) +
  geom_text(aes(label = abs(Difference),
                y = city,
                x = Position),
            nudge_y = 0.2) +
  theme_classic(base_size = 12) +
  labs(x = "Car owner)")+ 
  theme(legend.position = "top") +
  scale_color_brewer(palette = "Accent", direction = ) +
  theme(axis.ticks.y = element_line(colour = c("transparent",
                                               rep("black", 12))))

需求 #

希望按1992-1999年汽车拥有率的差值大小对图中城市排序，顺序为：A（差值59.9个百分点）、B、F、T、C。

修改后的代码 #

要实现按差值排序，核心是将city转换为有序因子，指定按差值降序排列的层级。修改后的完整代码如下：

library(dplyr)
library(ggplot2)

# 先计算每个城市的差值并排序，获取城市顺序
city_order <- df %>%
  group_by(city) %>%
  summarize(diff = abs(car_owner[year == "1999"] - car_owner[year == "1992"])) %>%
  arrange(desc(diff)) %>%
  pull(city)

# 将city转换为有序因子，指定排序层级
df$city <- factor(df$city, levels = city_order)

# 绘图（同时修正原代码中的小错误）
df %>%
  group_by(city) %>%
  mutate(Difference = car_owner - lead(car_owner),
         Position = car_owner - (0.5 * Difference)) %>%
  ggplot(aes(x = car_owner, y = city))+
  geom_line(aes(group = city)) +
  geom_point(aes(color = year), size=3) +
  geom_text(aes(label = abs(Difference),
                y = city,
                x = Position),
            nudge_y = 0.2) +
  theme_classic(base_size = 12) +
  labs(x = "汽车拥有率(%)", y = "城市")+ 
  theme(legend.position = "top") +
  scale_color_brewer(palette = "Accent", direction = 1) +
  theme(axis.ticks.y = element_line(colour = "black"))

关键修改说明 #

1. 生成城市排序：通过summarize计算每个城市的汽车拥有率差值，按差值降序排列后提取城市名称，得到目标排序顺序。
2. 转换为有序因子：将原数据中的city列转换为因子，指定levels为上面得到的排序顺序，ggplot会自动按这个顺序渲染y轴。
3. 修正原代码问题：补全scale_color_brewer的direction参数值为1，修正x轴标签的语法错误，简化axis.ticks.y的颜色设置。

内容的提问来源于stack exchange，提问作者Santiago99