How to fit a line in geom_point(), to estimate the overall trend for each plot.

Each line represents one patient with baseline measurement (red dot) and a second measurement (blue dot).

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") 

CodePudding user response：

One option would be to first reshape you data to long format which allows to simplify your plotting code a bit and makes it easy to add a tend line via geom_smooth.

Using some fake random example data:

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'