1
votes

I have a question about legends in ggplot2. I managed to plot two lines and two points in the same graph and want to add a legend with the two colors used. This is the code used

P <- ggplot() + geom_point(data = data_p,aes(x = V1,y = V2),shape = 0,col = "#56B4E9") + geom_line(data = data_p,aes(x = V1,y = V2),col = "#56B4E9")+geom_point(data = data_p,aes(x = V1,y = V3),shape = 1,col = "#009E73") + geom_line(data = data_p,aes(x = V1,y = V3),col = "#009E73")

and the output is enter image description here

I try to use scale_color_manual and scale_shape_manual and scale_line_manual,but they don't work .

P + scale_color_manual(name = "group",values = c('#56B4E9' = '#56B4E9','#009E73' = '#009E73'),
                   breaks = c('#56B4E9','#009E73'),labels = c('B','H')) +

I want it like this

Here is the simple data if it can help you.

5   0.49216 0.45148  
10  0.3913  0.35751  
15  0.32835 0.30361

data_p

1

1 Answers

0
votes

I would approach this problem in two steps.

Generally, to get stuff in the guides, ggplot2 wants you to put "aesthetics" like colour inside the aes() function. I typically do this inside the ggplot() rather than individually for each "geom", especially if everything kind of makes sense in a single dataframe.

My first step would be to remake your dataframe slightly. I would use the package tidyr (part of the tidyverse, like ggplot2, which is really nice for reformatting data and worth learning as you go), and do something like this

#key is the new variable that will be your color variable
#value is the numbers that had been in V2 and V3 that will now be your y-values
data_p %>% tidyr::gather (key = "color", value = "yval", V2, V3) 

#now, I would rewrite your plot slightly
P<-(newdf %>% ggplot(aes(x=V1,y=yval, colour=color))

#when you put the ggplot inside parentheses, 
#you can add each new layer on its own line, starting with the "+"
                 + geom_point()
                 + geom_line()
                 + scale_color_manual(values=c("#56B4E9","#009E73"))

#theme classic is my preferred look in ggplot, usually
                 + theme_classic()
)