How to merge colour and shape?

Question

My troubles started when I had a variable with more than 6 values because that is the current maximum value for the scale_shape function in ggplot2.

Due to that problem

Answer 1

Further trick: If you give either legend a name, you must give them both the same name. If you give only one legend a name, ggplot will separate the legends again. Amending kohske's example:

plot1 <- ggplot(dataf, aes(x=Density, y=Growth, colour=Municipality,
        shape=Municipality)) + geom_point(size=3)

plot2 <- plot1 + scale_colour_discrete() + 
scale_shape_manual(values=as.numeric(dataf$Municipality))

plot2

plot3 <- plot1 + scale_colour_discrete('City') + 
scale_shape_manual(values=as.numeric(dataf$Municipality))

plot3

plot4 <- plot1 + scale_colour_discrete('City') + 
scale_shape_manual('City',values=as.numeric(dataf$Municipality))

plot4

Answer 2

What about using scale_shape_manual()? If I understood your question correctly, you don't really need to differentiate by both colour and shape and would prefer shape, right?

ggplot(dataf, aes(x=Density, y=Growth)) + 
  geom_point(aes(shape = Municipality)) +
  scale_shape_manual(values = 1:11)

produces:

Answer 3

here is an example:

plot1 <- ggplot(dataf, aes(x=Density, y=Growth, colour=Municipality,
        shape=Municipality))
plot1 <- plot1 + geom_point(size=3)
plot1 <- plot1 + scale_colour_discrete() + 
scale_shape_manual(values=as.numeric(dataf$Shape))
plot1

if you need filled shapes, then replace with

scale_shape_manual(values=c(16, 17, 15, 3, 7, 8)[as.numeric(dataf$Shape)])

the tricks are:

1. use same variable for colour and shape aes (Municipality)
2. use scale_shape_manual and make mapping of breaks (here, Municipality) and value (here, dataf$Shape)
3. you need numeric variable instead of factor for values of scale_shape_manual