R：ggplot2 中 geom_point() 的平均趨勢線來估計總體趨勢？

Question

如何在 geom_point() 中擬合一條線，以估計每個 plot 的總體趨勢。

每條線代表一名具有基線測量（紅點）和第二次測量（藍點）的患者。

p <- ggplot(data = df, aes(x=days, y=temp_1)) +
  geom_point(aes(x=days, y=temp_1), color="red", show.legend = TRUE) +
  geom_point(aes(x=day, y=temp_2), color="blue", show.legend = TRUE)  +
  scale_color_manual(name="Measurements", breaks=c("temp_1", "temp_2"), values = c("red", "blue")) +
  xlim(0, 20) + ylim(70, 100) +
  geom_segment(aes(x=days, y=temp_1, xend=day, yend=temp_2))+
  facet_grid(~ treatment) +
  theme(legend.position = "bottom") 

Answer 1

一種選擇是首先將您的數據重塑為長格式，這可以稍微簡化您的繪圖代碼，並可以輕松地通過geom_smooth添加趨勢線。

使用一些假的隨機示例數據：

set.seed(123)

n <- 100
df <- data.frame(
  days = 0,
  day = sample(1:20, n, replace = TRUE),
  temp_1 = sample(70:100, n, replace = TRUE),
  temp_2 = sample(70:100, n, replace = TRUE),
  treatment = sample(c("yes", "no"), n, replace = TRUE)
)

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

df <- df %>%
  rename(day_1 = days, day_2 = day) %>%
  mutate(id = seq(nrow(.))) %>%
  pivot_longer(-c(id, treatment), names_to = c(".value", "time"), names_sep = "_")

ggplot(data = df, aes(x = day, y = temp)) +
  geom_point(aes(color = time), show.legend = TRUE) +
  geom_line(aes(group = id)) +
  geom_smooth(method = "lm", se = FALSE) +
  scale_color_manual(name = "Measurements", values = c("red", "blue")) +
  xlim(0, 20) +
  ylim(70, 100) +
  facet_grid(~treatment) +
  theme(legend.position = "bottom")
#> `geom_smooth()` using formula 'y ~ x'