Merge legends of separate geom layers

2019-03-06 14:06发布

I am plotting point data from two separate layers on a map and would like to display the information of both layers in one legend.

Here's some code to illustrate the problem:

set.seed(42)  
lat1 <- rnorm(10, 54, 12)
long1 <- rnorm(10, 44, 12)
val1 <- rnorm(10, 10, 3)
lat2 <- rnorm(10, 54, 12)
long2 <- rnorm(10, 44, 12)
val2 <- rnorm(10, 10, 3)

df1 <- as.data.frame(cbind(long1, lat1, val1))
df2 <- as.data.frame(cbind(long2, lat2, val2))

library(ggplot2)
library(scales)

f <- ggplot() +
     geom_point(data=df1, aes(x=lat1, y=long1, size=val1, fill=val1), shape=21, alpha=0.6) +
     scale_size_continuous(range = c(2, 12), breaks=pretty_breaks(4)) +
     scale_fill_distiller(direction = -1, palette="RdYlBu", breaks=pretty_breaks(4)) +     
     guides(fill = guide_legend(), size = guide_legend()) +
     theme_minimal()
p <- f + geom_point(data=df2, aes(x=lat2, y=long2, color="val2"), shape=17, size=3) +
     scale_color_manual(values="black",name="")
p

The best I can manage, is to create two separate legends and to then remove one of the legend titles. Ideally, both the filled circles and the black diamond would be part of the same legend called e.g. "Value" and the black diamond would read e.g. "NA". Any help is much appreciated!

1条回答
三岁会撩人
2楼-- · 2019-03-06 15:12

We have to plot two different legends, but move them close to each other (using negative value in legend.spacing.y requires ggplot2_3.0.0). This approach creates another problem - two legends doesn't align, therefore we have to plot another set of diamonds (larger to match non-diamonds, but invisible alpha = 0)

ggplot() +
    geom_point(data = df1, 
               aes(lat1, long1, size = val1, fill = val1), 
               shape = 21, alpha = 0.6) +
    geom_point(data = df2, 
               aes(lat2, long2, color = "val2"), 
               shape = 17, size = 3) +
    geom_point(data = df2, 
               aes(lat2, long2, color = "val2"), 
               shape = 17, size = 11.5, alpha = 0) +
    scale_size_continuous(range = c(2, 12), breaks = pretty_breaks(4)) +
    scale_fill_distiller(direction = -1, palette = "RdYlBu", breaks = pretty_breaks(4)) +     
    scale_color_manual(values = "black", name = "Value\n") +
    labs(fill = NULL,
         size = NULL) +
    guides(fill = guide_legend(), 
           size = guide_legend(),
           color = guide_legend(order = 1)) +
   theme_minimal() +
   theme(legend.spacing.y = unit(-0.4, "cm"))

enter image description here

查看更多
登录 后发表回答