How to show values in ggplot instead of geom-CodePudding

I have a dataset that looks as follows:

cluster <- c(rep(c(1:4), 2))
score <- c(1.3, 7.2, 4.9, 7.5, 6.8, 4.1, 9.9, 5.8)
x_axis <- c(rep("indicator1", 4), rep("indicator2", 4))

dt <- data.table(cluster, score, x_axis)

and I am plotting the score for each indicator by cluster:

ggplot()  
  geom_point(data=dt, aes(x=x_axis, y=score))

I would like to do one of the following:

Replace the dots in the plot with cluster number (or put the number next to the dot), or
Use different shapes/color for each cluster and add a legend below the plot, indicating to which cluster does each shape correspond to.

How can I do this?

CodePudding user response：

Here's both options. To use shape for the clusters, pass factor(cluster) to the shape aesthetic of geom_point.

ggplot(dt, aes(x_axis, score))  
  geom_point(aes(shape = factor(cluster)), size = 4)  
  theme_bw(base_size = 16)  
  labs(shape = 'Cluster')

If you just want labels at each point, you don't need the shape aesthetic, but you might want to draw in some white circles behind your text:

ggplot(dt, aes(x_axis, score))  
  geom_point(size = 8, shape = 21, fill = 'white')  
  geom_text(aes(label = cluster), size = 5)  
  theme_bw(base_size = 16)

If you're writing to a format where colors are permissable, I think the clearest way to label the clusters would be to map them to the color aesthetic:

ggplot(dt, aes(x_axis, score))  
  geom_point(aes(color = factor(cluster)), size = 8)  
  theme_bw(base_size = 16)  
  scale_color_brewer(palette = 'Set1')  
  labs(color = 'Cluster')