How to create a linetype legend in ggplot with different dataframes-CodePudding

I have been struggling with creating a legend in my ggplot for some time now, and I cannot find any answer that works.

This is the stripped down version of my ggplot:

ggplot()   
  geom_smooth(data = mydf1, aes(x, predicted, linetype = 1), method = "lm", linetype = 1, colour = "black")  
  geom_smooth(data = mydf2, aes(x, predicted, linetype = 2), method = "lm", linetype =  2, colour = "black")  
  geom_smooth(data = mydf3, aes(x, predicted, linetype = 3), method = "lm", linetype =  3, colour = "black")  
  theme_classic()

As you can notice, I am taking data from different dataframes (mydf1, mydf2, mydf3). Now I want to manually add a legend that specifies that the solid line is "Group 1", the longdashed line is "Group 2", and the dotted linetype is "Group 3". However, whatever I try, no legend appears in my ggplot. I included linetype within the aes(), tried everything I could think of regarding scale_linetype_manual(), and I've been looking around to find solutions, but no legend pops up.

Am I missing something obvious? I just need a small legend to the side stating what the different linetypes mean.

My data is the following:

mydf1 <- data.frame(x = c(seq(-1,1, 0.2)),
                    predicted = c(-0.27066438, -0.23568714, -0.20070991, -0.16573267, -0.13075543, -0.09577819, -0.06080095, -0.02582371,  0.00915353,  0.04413077,  0.07910801))

mydf2 <- data.frame(x = c(seq(-1,1, 0.2)),
                    predicted = c(-0.39806988, -0.34348641, -0.28890295, -0.23431948, -0.17973602, -0.12515255, -0.07056909, -0.01598562,  0.03859784,  0.09318131,  0.14776477))

mydf3 <- data.frame(x = c(seq(-1,1, 0.2)),
                    predicted = c(-0.25520076, -0.22917917, -0.20315758, -0.17713600, -0.15111441, -0.12509282, -0.09907123, -0.07304964, -0.04702806, -0.02100647, 0.00501512))

Although any practical solution would be helpful, I am not neccesarily looking to combine the dataframes into a single one and redo the ggplot. Thanks!

CodePudding user response：

Bring the line linetype inside aes() and use lab() to give title for the legend.

ggplot()   
  geom_smooth(data = mydf1, aes(x, predicted, linetype = "Group 1"), method = "lm", colour = "black")  
  geom_smooth(data = mydf2, aes(x, predicted, linetype = "Group 2"), method = "lm", colour = "black")  
  geom_smooth(data = mydf3, aes(x, predicted, linetype = "Group 3"), method = "lm", colour = "black")  
  labs(linetype="Group")  
  theme_classic()

CodePudding user response：

Here is an alternative way using bind_rows:

library(dplyr)
library(ggplot2)
bind_rows(mydf1, mydf2, mydf3) %>% 
  mutate(group = as.integer(gl(n(), 11, n()))) %>% 
  ggplot(aes(x, predicted, linetype=factor(group))) 
    geom_smooth(color="black", method = "lm", se=F)  
  scale_linetype_discrete(name="group", 
                          breaks=c(1, 2, 3), 
                          labels = c("group1", "group2", "group3")) 
  theme_classic()