I have a data frame, with 5 columns and 51 rows like the one below:

I wanted to graph as many lines s are the columns to see the trend of each var over the years.

I tried with:

lt <- c("VAR2" = "twodash", "VAR3" = "dotted", "VAR4" = "dashed", "VAR5" = "solid")

ggp1 <- ggplot(DATAFRAME, aes(VAR1))         # Create ggplot2 plot
  geom_line(aes(y = VAR2, linetype = "twodash", group=1, ))  
  geom_line(aes(y = VAR3, linetype = "dotted", group=1))  
  geom_line(aes(y = VAR4, linetype = "dashed", group=1))   
  geom_line(aes(y = VAR5, linetype = "solid", group=1))  
  labs(y = "number of parties in government", x= "year", lintype = "Legend") 
  scale_linetype_manual(values = lt) 
ggp1 

But it doesn't work. How can I do?

thank you all for the help.

CodePudding user response：

This should do it:

library(ggplot2)
DATAFRAME <- data.frame(
  VAR1 = 1:20, 
  VAR2 = runif(20,0,1), 
  VAR3 = runif(20,1,2), 
  VAR4 = runif(20,2,3), 
  VAR5 = runif(20,3,4)
)


lt <- c("VAR2" = "twodash", "VAR3" = "dotted", "VAR4" = "dashed", "VAR5" = "solid")

ggplot(DATAFRAME, aes(VAR1))         # Create ggplot2 plot
  geom_line(aes(y = VAR2, linetype = "VAR2", group=1, ))  
  geom_line(aes(y = VAR3, linetype = "VAR3", group=1))  
  geom_line(aes(y = VAR4, linetype = "VAR4", group=1))   
  geom_line(aes(y = VAR5, linetype = "VAR5", group=1))  
  labs(y = "number of parties in government", x= "year", lintype = "Legend")  
  scale_linetype_manual(values = unname(lt) )

The trick here is that when you create aesthetics this way (e.g., linetype = "VAR2" or linetype = "twodash", ggplot2 will turn them into a factor and order the levels alphabetically. In your original code, the levels would be, in order, dashed, dotted, solid and twodash. What I did above was to put the levels in the intended order and then gave the vector of line-type values. Another catch is that ggplot wants that vector to be un-named.

You could also specify breaks in scale_linetype_manual to set the levels of the factor that is created in the background. Then, your code would work as well.

ggplot(DATAFRAME, aes(VAR1))         # Create ggplot2 plot
  geom_line(aes(y = VAR2, linetype = "twodash", group=1, ))  
  geom_line(aes(y = VAR3, linetype = "dotted", group=1))  
  geom_line(aes(y = VAR4, linetype = "dashed", group=1))   
  geom_line(aes(y = VAR5, linetype = "solid", group=1))  
  labs(y = "number of parties in government", x= "year", lintype = "Legend")  
  scale_linetype_manual(values = unname(lt), breaks=c("twodash", "dotted", "dashed", "solid"))

^{Created on 2022-04-12 by the}

CodePudding user response：

you can also transpose the data with pivot_wider() so you don't have to call geom_line multiple times. You then can feed in the new variable with all of your variables to linetype as well.

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

DATAFRAME <- data.frame(
  VAR1 = 1:20, 
  VAR2 = runif(20,0,1), 
  VAR3 = runif(20,1,2), 
  VAR4 = runif(20,2,3), 
  VAR5 = runif(20,3,4)
)
#transpose to long format except your year variable
df2 <- pivot_longer(DATAFRAME, cols =  paste0("VAR",2:5))

ggplot(df2)  
  geom_line(aes(x=VAR1 , y = value,  group =name, linetype = name))  
  labs(y = "number of parties in government", x= "year", lintype = "Legend")