ggplot2使用geom_encircle时如何移除NA值仅保留指定组环绕多边形

解决方法 #

问题根源是geom_encircle的na.rm参数仅对x、y列的缺失值生效，不会自动过滤fill映射列class的NA值，因此class为NA的行仍然会被识别为单独分组绘制多余的轮廓。

要实现保留全部散点、仅绘制class非空分组的环绕效果，只需要在geom_encircle层单独传入过滤掉class为NA的子集即可，不会影响geom_point层的全量数据展示，修改后代码如下：

x <- c(10, 12, 4, 18, 6, 9, 2, 2, 7, 23, 13, 13, 11, 6, 22)
y <- c(3, 2, 12, 15, 23, 20, 6, 21, 6, 8, 15, 19, 10, 18, 14)
group <- c("a", "b", "b", "b","b","b","b", "c", "c", "c","c","c","c", "d", "e")
class <- c(NA, "1", "1","1","1","1","1","2","2","2","2","2","2", NA, NA)

df <- as.data.frame(cbind(x,y,group,class))
df$x <- as.numeric(df$x)
df$y <- as.numeric(df$y)

library(ggplot2)
library(ggalt)

ggplot(df, aes(x, y)) +
  geom_point(aes(color = group)) +
  # 仅在geom_encircle层使用筛选后的数据
  geom_encircle(data = ~.x[!is.na(.x$class), ],
                aes(fill = class), s_shape = 1, expand = 0,
                alpha = 0.2, color = "black", show.legend = FALSE)

代码中~.x[!is.na(.x$class), ]是ggplot2的语法糖，表示基于全局传入的数据集df做筛选，也可以直接替换为df[!is.na(df$class), ]，二者效果完全一致。

内容的提问来源于stack exchange，提问作者Stavrum