Group geom_point with the geom_polygon

Question

My dataset:

Taxa    dn  dc
Cha    10.2 -20.4
Cha    10.7 -19.7
Cha    4.9  -21.0
Cha    5.4  -20.6
Cha    8.6  -21.2
Cha    8.0  -20.9
Cha    8.1  -21.3
Cha    6.9  -21.1
Cha    8.5  -21.1
Cha    9.1  -20.8
Hyd    6.6  -19.2
Hyd    10.2 -17.0
Hyd    9.7  -18.2
Hyd    8.1  -16.5
Hyd    8.8  -15.8
Hyd    8.7  -15.8
Hyd    7.6  -18.3
Hyd    8.9  -16.0
Hyd    8.4  -17.5
Hyd    9.8  -18.8
Hyd    8.3  -18.4
Scy    9.4  -20.1
Scy    9.1  -20.0
Scy    7.8  -20.2
Scy    9.1  -17.6
Scy    8.2  -19.8
Scy    9.4  -19.2
Scy    9.0  -20.1
Sip    5.7  -15.2
Sip    6.2  -18.6
Sip    5.6  -18.0
Sip    8.6  -17.6
Sip    4.8  -16.9
Sip    5.2  -15.4
Sip    1.9  -18.4

The code I use is:

library(ggplot2)
ggplot(mydata, aes(x=dC, y=dN, colour=Taxa, shape=Taxa))+
  geom_point(size=2, alpha=0.5)+
  geom_polygon(aes(fill=Taxa, group=Taxa))+
  theme(legend.position = "none")

I would like to plot the polygon group with "Taxa" in my data. However, it looks like the polygon connects each point.

What I want is like this one. How should I edit my codes?

Answer 1

To connect outer points in group and encircle ones that are within the group use geom_encircle function from ggalt package.

library(ggplot2)
library(ggalt)
ggplot(mydata, aes(dc, dn)) +
    geom_point(aes(color = Taxa)) +
    geom_encircle(aes(fill = Taxa), s_shape = 1, expand = 0,
                  alpha = 0.2, color = "black", show.legend = FALSE)

Use s_shape = 1 and expand = 0 to connect outer points, otherwise it will encircle with margins.

Answer 2

You can also calculate the convex hulls, them plot them down:

library(ggplot2)
library(plyr)

# some fake data:
mydata <- data.frame(Taxa = c('Cha','Cha','Cha','Cha','Cha','Cha','Hyd','Hyd','Hyd','Hyd','Hyd','Hyd'),
                     dn = c(10.2,10.7,4.9,5.4,8.6,8.0, 6.6,10.2,9.7,8.1,8.8,8.7),
                     dc =c(-20.4,-19.7,-21.0,-20.6,-21.2,-20.9,-19.2,-17.0,-18.2,-16.5,-15.8,-15.8))

# calculate convex hulls:
chulls <- ddply(mydata, .(Taxa), function(mydata) mydata[chull(mydata$dn, mydata$dc), ])

# plot them:
ggplot(data=mydata, aes(x=dn, y=dc, color=Taxa)) + geom_point() +
                   geom_polygon(data=chulls, aes(x=dn, y=dc, fill=Taxa, alpha=0.2)) 

Nice source here.