如何在Plotly箱线图中同时设置填充色与边框色属性

问题描述 #

我有一组层级化数据，需要生成分组箱线图，要求按group_a设置填充色，按group_c设置边框色。这个需求用ggplot2很容易实现（示例如下），但无法直接用Plotly原生函数实现。另外因为大数据集下ggplotly()性能太差，所以不考虑这种转换方式。

额外需求：如何在Plotly方案中手动指定group_c的边框颜色，实现类似ggplot2中scale_colour_manual()的功能。

数据准备 #

注：原代码中length(lot_ids)为笔误，已修正为length(group_b)

group_a <- 900:904
group_b <- paste0(rep(group_a, each = 3), '_', rep(c('01', '02', '03'), times = length(group_a)))
group_c <- rep(c("A", "B"), times = length(group_b))

# 修正原代码中的变量错误
group_date <- seq.Date(from = as.Date("2025-01-01"), by = "day", length.out = length(group_b))
group_date <- rep(group_date, each = 3)

x <- rep(runif(14), times = length(group_a) * 3 * 2)

# 创建数据框
df <- data.frame(
  group_a = rep(group_a, each = 3 * 2 * 14),
  group_b = rep(group_b, each = 14 * 2),
  group_date = rep(group_date, each = 2 * 14),
  group_c = group_c,
  x = x
)

ggplot2参考实现 #

library(ggplot2)
library(dplyr)

df %>%
  ggplot(aes(
    x = group_date,
    y = x,
    group = interaction(group_c, group_b),
    fill = factor(group_a),
    colour = group_c
  )) + 
  geom_boxplot()

Plotly原生实现方案 #

1. 自定义颜色映射 #

先手动定义group_c的边框色（对应ggplot2的scale_colour_manual），以及group_a的填充色：

library(plotly)

# 手动指定group_c的边框颜色
group_c_colors <- c("A" = "#1f77b4", "B" = "#ff7f0e")

# 手动指定group_a的填充色（也可使用默认色阶）
group_a_colors <- c("900" = "#aec7e8", "901" = "#ffbb78", "902" = "#98df8a", "903" = "#ff9896", "904" = "#c5b0d5")

2. 处理分组与颜色关联 #

为数据框添加交互分组列，并绑定对应的填充色和边框色：

df <- df %>%
  mutate(
    # 创建与ggplot一致的交互分组
    plot_group = interaction(group_c, group_b),
    # 匹配group_a对应的填充色
    fill_color = group_a_colors[as.character(group_a)],
    # 匹配group_c对应的边框色
    line_color = group_c_colors[group_c]
  )

3. 绘制分组箱线图 #

用Plotly原生函数实现分组效果，同时应用自定义颜色：

plot_ly(df, 
        x = ~group_date, 
        y = ~x, 
        split = ~plot_group,
        type = "box",
        fillcolor = ~fill_color,
        line = list(color = ~line_color),
        showlegend = FALSE) %>%
  layout(
    xaxis = list(title = "日期"),
    yaxis = list(title = "x值"),
    boxmode = "group" # 开启分组模式，让同日期的箱线图并排显示
  )

关键说明 #

• split = ~plot_group：实现和ggplot中group = interaction(group_c, group_b)完全一致的分组逻辑
• fillcolor = ~fill_color：将填充色与group_a绑定
• line = list(color = ~line_color)：将边框色与自定义的group_c颜色映射绑定
• boxmode = "group"：确保同一x轴位置的多个箱线图并排，还原ggplot的分组效果

内容的提问来源于stack exchange，提问作者Ali