连续数据分箱后保序：ggplot2自动整理坐标轴标签方法

模拟数据集代码 #

# Step 1 : Simulate Data

set.seed(123)
Hospital_Visits = sample.int(20,  5000, replace = TRUE)
Weight = rnorm(5000, 90, 10)

disease <- c("Yes","No")
disease <- sample(disease, 5000, replace=TRUE, prob=c(0.4, 0.6))
Disease <- as.factor(disease)

my_data = data.frame(Weight, Hospital_Visits, Disease)

my_data$hospital_ntile <- cut(my_data$Hospital_Visits, breaks = c(0, 5, 10, Inf), labels = c("Less than 5", "5 to 10", "More than 10"), include.lowest = TRUE)

数据处理代码 #

# Step 2: Data Manipulation:

my_data$weight_ntile <- cut(my_data$Weight, breaks = seq(min(my_data$Weight), max(my_data$Weight), by = (max(my_data$Weight) - min(my_data$Weight)) / 10), include.lowest = TRUE)


# Create a dataset for rows where hospital_ntile = 'Less than 5'
df1 <- subset(my_data, hospital_ntile == "Less than 5")

# Create a dataset for rows where hospital_ntile = '5 to 10'
df2 <- subset(my_data, hospital_ntile == "5 to 10")

# Create a dataset for rows where hospital_ntile = 'More than 10'
df3 <- subset(my_data, hospital_ntile == "More than 10")

avg_disease_rate_df1 <- tapply(df1$Disease == "Yes", df1$weight_ntile, mean)
avg_disease_rate_df2 <- tapply(df2$Disease == "Yes", df2$weight_ntile, mean)
avg_disease_rate_df3 <- tapply(df3$Disease == "Yes", df3$weight_ntile, mean)

avg_disease_rate_df1[is.na(avg_disease_rate_df1)] <- 0
avg_disease_rate_df2[is.na(avg_disease_rate_df2)] <- 0
avg_disease_rate_df3[is.na(avg_disease_rate_df3)] <- 0

#transform into dataset

names = names(avg_disease_rate_df1)
rate_1 = as.numeric(avg_disease_rate_df1)
rate_2 = as.numeric(avg_disease_rate_df2)
rate_3 = as.numeric(avg_disease_rate_df3)

# stack data
d1 = data.frame(class = "Less than 5", names = names, rate = rate_1)
d2 = data.frame(class = "5 to 10", names = names, rate = rate_2)
d3 = data.frame(class = "More than 10", names = names, rate = rate_3)

plot_data = rbind(d1, d2, d3)

绘图代码与问题 #

library(ggplot2)
ggplot(plot_data, aes(x=names, y=rate, group = class,  color=class)) + geom_point() + geom_line() +  theme_bw()

目前x轴标签顺序混乱，并非从小到大排列，希望将其调整为正确顺序。已查阅手动调整的方法，询问：ggplot2中是否存在可自动修正该顺序的选项？

解答 #

核心原因 #

weight_ntile是cut()生成的有序因子，本身带有从小到大的分箱顺序，但后续处理中（比如提取tapply结果的names并转为数据框列），这个顺序丢失，变成无序因子或字符型，ggplot默认按字符排序，导致顺序混乱。

自动修正的方法 #

不需要手动指定顺序，只要保留因子的原始顺序即可，以下是两种简便方案：

方案1：保留因子属性构建绘图数据

修改数据处理最后阶段的代码，让plot_data$names继承原因子的顺序：

# 替换原stack data部分的代码
# 直接使用原weight_ntile的因子类型，而非提取字符型names
d1 = data.frame(class = "Less than 5", names = df1$weight_ntile, rate = rate_1)
d2 = data.frame(class = "5 to 10", names = df2$weight_ntile, rate = rate_2)
d3 = data.frame(class = "More than 10", names = df3$weight_ntile, rate = rate_3)

plot_data = rbind(d1, d2, d3)
# 统一因子水平，避免因子集缺失导致的水平不一致
plot_data$names = factor(plot_data$names, levels = levels(my_data$weight_ntile))

此时ggplot会自动按照因子的原始分箱顺序排列x轴。

方案2：用fct_inorder()自动恢复顺序

如果plot_data$names已经是字符型，可借助forcats包（ggplot2依赖包，无需额外安装）的fct_inorder()函数，让ggplot按数据中首次出现的顺序排列：

library(ggplot2)
library(forcats)
ggplot(plot_data, aes(x=fct_inorder(names), y=rate, group = class,  color=class)) + 
  geom_point() + 
  geom_line() +  
  theme_bw()

fct_inorder()会将字符列转为因子，顺序与数据中首次出现的顺序一致，而你的分箱是按weight从小到大生成的，因此这个顺序就是正确的。

可选：简化数据处理流程

用dplyr可以更高效地处理数据，同时自动保留因子顺序，无需额外调整：

library(dplyr)
plot_data = my_data %>%
  group_by(hospital_ntile, weight_ntile) %>%
  summarise(rate = mean(Disease == "Yes", na.rm = TRUE), .groups = "drop") %>%
  rename(class = hospital_ntile, names = weight_ntile) %>%
  mutate(rate = replace(rate, is.na(rate), 0))

生成的plot_data$names直接保留了weight_ntile的有序因子属性，绘图时x轴会自动按正确顺序排列。

内容的提问来源于stack exchange，提问作者stats_noob