Add legend for multiple regression lines from different datasets to ggplot

Question

I have loaded 4 databases into R, each of them have 2 columns: "Rating" and "Temperature". I want to create a graph with 4 regression lines (without data points inside the plot) of Ratings and Temperature. My code is:

y1<-Database1$`Rating1'
y2<-Database2$'Rating2'
y3<-Database3$`Rating3`
y4<-Database4$`Rating4`

x1<-Database1$`Mean Temperature (°C)`
x2<-Database2$`Mean Temperature (°C)`
x3<-Database3$`Mean Temperature (°C)`
x4<-Database4$`Mean Temperature (°C)`

reg1 <- lm(y1 ~ x1, data=Database1)
reg2 <- lm(y2 ~ x2, data=Database2)
reg3 <- lm(y3 ~ x3, data=Database3)
reg4 <- lm(y4 ~ x4, data=Database4)

equation1=function(x){coef(reg1)[2]*x+coef(reg1)[1]}
equation2=function(x){coef(reg2)[2]*x+coef(reg2)[1]}
equation3=function(x){coef(reg3)[2]*x+coef(reg3)[1]}
equation4=function(x){coef(reg4)[2]*x+coef(reg4)[1]}

library(ggplot2)
ggplot(Database1,aes(`Mean Temperature (°C)`,`Rating1`))+
  ggtitle("Sub-ratings vs Temperature (°C)") +
  xlab("Temperature (°C)") + ylab("Sub-ratings")+
  stat_function(fun=equation1,geom="line",color="skyblue",size=1)+
  stat_function(fun=equation2,geom="line",color="orangered",size=1)+
  stat_function(fun=equation3,geom="line",color="turquoise",size=1)+
  stat_function(fun=equation4,geom="line",color="royalblue",size=1)+
  theme(
    plot.title = element_text(size=12,hjust = 0.5),
    axis.title.x = element_text(size=11),
    axis.title.y = element_text(size=11)
  )+ylim(2,5)

I have created a graph like this.

I hope to add a legend for 4 lines to explain which color is for which Rating. Would be better to place the legend outside.

Thank you!

Answer 1

Legends are created for any aesthetic that is placed in aes(). Typically, they are created by plotting from one data frame and just supplying aes(color= with a column from that data frame. In the absence of a column connecting the geoms for which you want to create a legend (which is your case), you can utilize the knowledge that ggplot will make a legend whenever you put some differentiating aesthetic (like color=) within aes().

To add your legend, add character strings within aes(color= for each of your points that will serve as the label (key value label) in the legend. Then, you can use scale_color_manual to define the name of the legend as well as the colors themselves.

The following is an example, where I'm adding 3 points to an existing plot and I want to call out the color of each.

df <- data.frame(x=1:10, y=1:10)
ggplot(df, aes(x, y)) +
  geom_point() + theme_bw() +
  geom_point(data=data.frame(x=2,y=4), size=3, aes(color='point1')) +
  geom_point(data=data.frame(x=3,y=9), size=3, aes(color='point2')) +
  geom_point(data=data.frame(x=1,y=6), size=3, aes(color='point3')) +
  scale_color_manual(
    name='My Points',
    values=c('blue', 'orangered', 'green3')
  )

If you are having trouble assigning specifically one legend key to a specific color, you can also feed a list into values= for scale_color_manual that defines 'key.name'='color.name' for each item.