R: how to ggplot differently colored/shaped points-CodePudding

set.seed(3)
mydat <- data.frame(ref = rnorm(5), mars = rnorm(5), saturn = rnorm(5), time = c(0, 0.5, 1, 1.5, 2))
> mydat
         ref        mars     saturn time
1  0.9006247  0.70551551  0.7865069  0.0
2  0.8517704  1.30035799 -0.3104631  0.5
3  0.7277152  0.03825201  1.6988848  1.0
4  0.7365021 -0.97928377 -0.7945937  1.5
5 -0.3521296  0.79376123  0.3484377  2.0

I have a dataset called mydat, where I would like to make the following plot with a diagonal line for ref. Then for mars and saturn, I would like to plot them with different colors and symbols. I can do this using plot just fine.

plot(mydat$ref, mydat$ref, type = "l",
     xlab = "Reference", ylab = "Actual measurement", ylim = c(-1, 2))
points(mydat$ref, mydat$mars, pch = 15, col = "blue")
points(mydat$ref, mydat$saturn, pch = 18, col = "green")

I would like to plot this in ggplot. However, I tried the following but couldn't get the same plot:

library(ggplot2)
library(reshape2)
melt_mydat <- reshape2::melt(mydat, id.vars = "time")
ggplot(data = melt_mydat, mapping = aes(x = value, y = value, color = as.factor(variable), group = as.factor(variable)))    geom_point()

CodePudding user response：

Maybe you want something like this:

library(ggplot2)
ggplot(mydat, aes(x = ref))  
  geom_point(aes(y = mars), color = "blue")  
  geom_point(aes(y = saturn), color = "green")  
  geom_line(aes(y = ref))  
  labs(x = "Reference", y = "Actual measurement")  
  theme_minimal()

Output:

CodePudding user response：

You just need to keep ref as its own column when using melt (i.e. specify it as one of id.vars). Then we can use its value for the x-axis in the plot:

set.seed(3)
mydat <- data.frame(ref = rnorm(5), mars = rnorm(5), saturn = rnorm(5), time = c(0, 0.5, 1, 1.5, 2))

library(ggplot2)
library(reshape2)
melt_mydat <- reshape2::melt(mydat, id.vars = c("time","ref"))

ggplot(data = melt_mydat, aes(x = ref, y = value, color = variable))  
  geom_point()  
  geom_abline(slope = 1)  
  scale_color_manual(values = c("blue", "green"))  
  labs(x = "Reference", y = "Actual measurement")