创建一个带有百分比的ggplot

Posted 2023-02-16

【中文标题】创建一个带有百分比的ggplot

【英文标题】：Create a ggplot with percentages

【发布时间】：2021-10-14 14:49:05

【问题描述】：

我想制作一个 ggplot（线图），在 x 轴（y 轴）上显示每个时间步长的值（A、B、C、D）的百分比。当我融化我的数据框时，我只有两列；如何计算百分比？

数据结构：

样本数据：

structure(list(`09:20` = c("A", "A", "A", "A", "B", "C", "D"), 
    `09:30` = c("B", "B", "B", "A", "A", "C", "C"), `09:40` = c("C", 
    "C", "C", "B", "B", "D", "D"), `09:50` = c("A", "A", "A", 
    "D", "D", "D", "D")), class = c("spec_tbl_df", "tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -7L), spec = structure(list(
    cols = list(`09:20` = structure(list(), class = c("collector_character", 
    "collector")), `09:30` = structure(list(), class = c("collector_character", 
    "collector")), `09:40` = structure(list(), class = c("collector_character", 
    "collector")), `09:50` = structure(list(), class = c("collector_character", 
    "collector"))), default = structure(list(), class = c("collector_guess", 
    "collector")), skip = 1L), class = "col_spec"))

【参考方案1】：

在重塑为长格式后，使用例如group_by + summarise 计算每个时间和值的计数，然后按时间和绘图计算百分比：

library(dplyr)
library(tidyr)
library(ggplot2)

d1 <- d %>% 
  pivot_longer(everything(), names_to = "time", values_to = "value") %>% 
  group_by(time, value) %>% 
  summarise(n = n()) %>%  
  mutate(pct = n / sum(n)) %>% 
  ungroup()
#> `summarise()` has grouped output by 'time'. You can override using the `.groups` argument.

ggplot(d1, aes(time, pct, color = value, group = value)) +
  geom_line() +
  scale_y_continuous(labels = scales::percent)