ggplot箱线图添加观测权重的实现方法求助

需求：基于包含观测权重pesofit的数据集，绘制按地区（nome_reg）划分的工资（tpens）加权箱线图，但在ggplot的aes中添加weight=pesofit后出现警告，无法实现加权效果。

数据集样例 #

> head(final2, 20)
   nquest nord ireg staciv etalav acontrib       nome_reg tpens   pesofit
1     173    1   18      3     25       35       Calabria  1800 0.3801668
2    2886    1   13      1     26       35        Abruzzo  1211 0.2383701
3    2886    2   13      1     20       42        Abruzzo  2100 0.2383701
4    5416    1    8      3     16       30 Emilia Romagna   700 0.8819879
5    7886    1    9      1     22       35        Toscana  2000 1.2452078
6   20297    1    5      1     14       39         Veneto  1200 1.6694498
7   20711    2    4      1     15       37       Trentino  2000 3.3746801
8   22169    1   15      4     40        5       Campania   600 1.6875562
9   22276    1    8      2     18       37 Emilia Romagna  1200 2.1782894
10  22286    1    8      1     15       19 Emilia Romagna   850 3.0333999
11  22286    2    8      1     15       35 Emilia Romagna   650 3.0333999
12  22657    1   16      1     25       40         Puglie  1400 0.3616937
13  22657    2   16      1     26       36         Puglie  1500 0.3616937
14  23490    1    5      2     23       36         Veneto  1400 0.9763965
15  24147    1    4      1     26       35       Trentino  1730 1.2479984
16  24147    2    4      1     18       45       Trentino  1600 1.2479984
17  24853    1   11      1     18       38         Marche  2180 0.3475683
18  27238    1   12      1     16       31          Lazio  1050 3.6358952
19  27730    1   20      1     15       37       Sardegna  1470 0.7232677
20  27734    1   20      1     16       45       Sardegna  1159 0.6959107

变量说明 #

• nquest：家庭编码
• nord：家庭成员编码
• nome_reg：居住地区
• tpens：个人工资
• pesofit：观测权重

原普通箱线图代码（可正常运行） #

final2 %>%
  filter(nome_reg == "Piemonte"| 
         nome_reg == "Valle D'Aosta" | 
         nome_reg == "Lombardia" | 
         nome_reg == "Liguria"
        ) %>%
          ggplot(aes( x = factor(nome_reg, 
                      levels=c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria")), 
                      y = tpens , fill = nome_reg ))+
  geom_boxplot(varwidth = TRUE) 

尝试加权后的代码及警告 #

尝试添加weight=pesofit后代码：

final2 %>%
  filter(nome_reg == "Piemonte"| 
         nome_reg == "Valle D'Aosta" | 
         nome_reg == "Lombardia" | 
         nome_reg == "Liguria") %>%
  ggplot(aes( x = factor(nome_reg, levels=c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria")), 
               y = tpens , fill = nome_reg, weight = pesofit ))+
  geom_boxplot(varwidth = TRUE)

返回警告：

Warning message:
The following aesthetics were dropped during statistical transformation: weight
i This can happen when ggplot fails to infer the correct grouping structure in the data.
i Did you forget to specify a `group` aesthetic or to convert a numerical variable into a factor?

核心原因：ggplot2默认的stat_boxplot不支持weight美学参数，无法直接通过添加weight实现加权箱线图，需要手动计算加权分位数或使用支持加权的扩展包。

方法一：手动计算加权分位数后绘图 #

使用Hmisc包的wtd.quantile函数计算各组的加权分位数，再用ggplot绘制箱线图：

1. 先安装并加载所需包：

install.packages("Hmisc")
library(dplyr)
library(ggplot2)
library(Hmisc)

2. 过滤数据并计算加权统计量：

weighted_box_data <- final2 %>%
  filter(nome_reg %in% c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria")) %>%
  mutate(nome_reg = factor(nome_reg, levels = c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria"))) %>%
  group_by(nome_reg) %>%
  summarize(
    # 计算加权分位数
    ymin = wtd.quantile(tpens, weights = pesofit, probs = 0),
    lower = wtd.quantile(tpens, weights = pesofit, probs = 0.25),
    middle = wtd.quantile(tpens, weights = pesofit, probs = 0.5),
    upper = wtd.quantile(tpens, weights = pesofit, probs = 0.75),
    ymax = wtd.quantile(tpens, weights = pesofit, probs = 1),
    # 计算加权样本量（用于varwidth）
    n = sum(pesofit)
  )

3. 绘制加权箱线图：

ggplot(weighted_box_data, aes(x = nome_reg, fill = nome_reg)) +
  geom_boxplot(
    aes(ymin = ymin, lower = lower, middle = middle, upper = upper, ymax = ymax),
    stat = "identity",
    varwidth = TRUE
  ) +
  labs(y = "tpens")

方法二：使用ggstatsplot包直接绘制加权箱线图 #

ggstatsplot包的ggboxplot函数原生支持weight参数，无需手动计算：

1. 安装并加载包：

install.packages("ggstatsplot")
library(ggstatsplot)
library(dplyr)

2. 绘制加权箱线图：

final2 %>%
  filter(nome_reg %in% c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria")) %>%
  mutate(nome_reg = factor(nome_reg, levels = c("Piemonte", "Valle D'Aosta", "Lombardia", "Liguria"))) %>%
  ggboxplot(
    x = "nome_reg",
    y = "tpens",
    fill = "nome_reg",
    weight = "pesofit",
    varwidth = TRUE
  )

内容的提问来源于stack exchange，提问作者io_boh