如何修改ggplot+ggwordcloud词云图例中的文本内容

问题背景 #

使用TidyTuesday数据集（统计《足球教练》中Roy Kent每集说fuck的次数）制作词云，首次使用ggwordcloud包时，发现词云图例中的文本显示为NA，需要修改为"FUCK"这类指定内容。

原始代码 #

环境配置 #

library(ggplot2)
library(ggtext)
library(ggwordcloud)
library(showtext)
library(tidyverse)

数据加载 #

tuesdata <- tidytuesdayR::tt_load(2023, week = 39)
fuck <- tuesdata$richmondway

绘图数据处理 #

fuck_plot <- fuck |>
              select("Season_Episode", "F_count_RK", "Dating_flag", "Coaching_flag") |>
              mutate(word = "FUCK")

字体设置 #

font_add_google("Oswald", "os")
showtext_auto()
plot_font <- "os"

原始绘图代码 #

fp <- fuck_plot |>
        ggplot(aes(x = 0, y = 0, label = word, size = F_count_RK, color = Dating_flag)) +
          geom_text_wordcloud_area(
            aes(color = Dating_flag),
            eccentricity = .5,
            seed = 289,
            tstep = 0.04,
            family = "os",
            fontface = "bold",
            show.legend = TRUE
          ) +
          scale_size_area(max_size = 40) +
          scale_color_manual(
            labels = c("Hell No", "Hell Yea"), 
            values = c("black", "#e0218a"), 
            name = "Is Rick Dating Keeley?"
          ) +
          coord_cartesian(
            expand = FALSE
          )  +
          guides(size = "none") +
          theme_void() +
          theme(
            plot.margin = margin(0, 0, 0, 0),
            legend.position = c(0.5, 0.75)
          )

解决方案 #

问题根源是geom_text_wordcloud_area的图例默认使用label映射的内容，但所有数据的word值相同，导致图例文本被识别为NA。通过在guides中自定义颜色图例的符号文本即可解决：

修改后的绘图代码 #

fp <- fuck_plot |>
        ggplot(aes(x = 0, y = 0, label = word, size = F_count_RK, color = Dating_flag)) +
          geom_text_wordcloud_area(
            eccentricity = .5,
            seed = 289,
            tstep = 0.04,
            family = "os",
            fontface = "bold",
            show.legend = TRUE
          ) +
          scale_size_area(max_size = 40) +
          scale_color_manual(
            labels = c("Hell No", "Hell Yea"), 
            values = c("black", "#e0218a"), 
            name = "Is Rick Dating Keeley?"
          ) +
          coord_cartesian(
            expand = FALSE
          )  +
          guides(
            size = "none",
            # 强制替换图例中的文本为指定内容
            color = guide_legend(override.aes = list(label = "FUCK"))
          ) +
          theme_void() +
          theme(
            plot.margin = margin(0, 0, 0, 0),
            legend.position = c(0.5, 0.75)
          )

关键修改说明 #

• 添加color = guide_legend(override.aes = list(label = "FUCK"))到guides中，直接替换颜色图例的文本内容，将NA改为"FUCK"。
• 移除了geom_text_wordcloud_area中重复的aes(color = Dating_flag)，因为顶层ggplot(aes())已经包含该映射，重复定义无必要。

内容的提问来源于stack exchange，提问作者ZahlenZ