如何在不同方面为 ggplot stat_smooth 设置不同的分组美学

如何解决如何在不同方面为 ggplot stat_smooth 设置不同的分组美学

我有以下数据：

structure(list(patient_code = c("BD1200","BD1200","BD1402","CA1601","CS1701","CE1801","DE1511","LM1400","MJ1402","PP1511","RA1401","RC1400","SO1701","SS1801","SS1801"
),age = c(91,91,73,56,46,38,61,71,76,67,74,41,40,40),sex = c("M","M","F","F"),height = c(170,170,156,180,162,169,165,153,185,168,173,158,182,165),weight = c(72,72,53,106,69,66,52,45,110,92,85,96.5,67),bmi = c(24.9134948096886,24.9134948096886,21.7784352399737,32.716049382716,26.291723822588,23.1084345786212,19.1000918273646,19.2233756247597,32.1402483564646,32.5963718820862,28.4005479635136,29.6426854670726,29.1329549571308,24.6097337006428,24.6097337006428),variable = c("muscle_fat_mm3","muscle_ff_from_segmentation","muscle_mm3","subcut_fat_mm3","muscle_fat_mm3","subcut_fat_mm3"),value = c(478569.3359375,45.600405239292,4169956.0546875,1476489.2578125,376596.6796875,31.5671469350962,2521455.078125,2642592.7734375,604804.6875,59.1302753496503,6097094.7265625,3302670.8984375,572958.984375,11.596700038243,3446499.0234375,3746103.515625,185146.484375,35.0908397071678,3897255.859375,3434550.78125,310375.9765625,31.8352193509615,3090175.78125,1303105.46875,407695.3125,23.0094809877622,2032436.5234375,1430712.890625,1012875.9765625,46.6928064903846,4884965.8203125,2261372.0703125,796171.875,54.2822574300699,4961044.921875,2377094.7265625,610234.375,25.5995147235577,3695327.1484375,1061621.09375,421166.9921875,35.237383631993,3163217.7734375,3767353.515625,453144.53125,16.9699877076049,6104433.59375,3296123.046875,404785.15625,9.00698576813811,3238256.8359375,3941044.921875)),row.names = c(NA,-52L),class = c("tbl_df","tbl","data.frame"))

如果我使用分面为 variable 和 BMI 中的每个变量的关联创建散点图，显示按 sex 分组的拟合线如下：

library(ggplot2)
ggplot(demo_index,aes(x = bmi,y = value,color = sex)) +
  geom_point(size = 1.5) +
  stat_smooth(aes(group = sex),method = "lm",se = FALSE,size = 1.5) + 
  facet_wrap(~variable,scales = "free") +
  ggtitle("Age associations") + 
  theme_bw() + 
  theme(panel.grid.minor = element_blank(),text = element_text(size = 10),plot.title = element_text(hjust = 0.5),axis.title.y = element_blank())

我得到了以下情节

很明显，虽然某些关联更适合将男性和女性分开，但某些其他关联在没有分组的情况下更适合。 我的问题是：ggplot 有没有办法像上面那样做一个分面散点图，其中 stat_smooth 的分组变量在分面之间变化？

解决方法

您不需要更改每个方面的分组变量，您只需要使用一个分组变量，该分组变量在该方面具有您想要的组。

library(tidyverse)

demo_index %>% 
  mutate(grp = if_else(str_starts(variable,"muscle_ff"),"both",sex)) %>% 
  ggplot(aes(x = bmi,y = value)) +
  geom_point(aes(color = sex),size = 1.5) +
  stat_smooth(aes(color = grp,group = grp),method = "lm",se = FALSE,size = 1.5) + 
  facet_wrap(~variable,scales = "free") +
  ggtitle("Age associations") + 
  theme_bw() + 
  theme(panel.grid.minor = element_blank(),text = element_text(size = 10),plot.title = element_text(hjust = 0.5),axis.title.y = element_blank())

我确实稍微调整了 aes(color) 映射，以便正确显示图例。

注意：对于您所在领域以外的人来说，哪些组应该一起建模并不是很明显，所以我做了我最好的猜测..