Dealing with points in sf data

[eric-pedersen]

Do we want to allow for points and develop MRFs for them?

Our best idea so far is to use a Voronoi tessellation to turn a given point set into a polygon set, but with non-convex point sets, we get odd issues. Basically, points that are far apart but at the edge of the area can be connected by this.

library(sf)
library(ggplot2)
set.seed(1)
#simulate some new data that falls in a crescent shape
new_data = data_frame(x = rnorm(600),y=rnorm(600))%>%
  filter(x^2 + y^2 <=1)%>%
  filter((x-0.2)^2 + (y-0.2)^2 >0.6)%>%
  st_as_sf(coords= c("x","y"))

ggplot(new_data) + geom_sf()

#Create a Voronoi tesselation that we can get adjacency from
new_voronoi <- st_combine(new_data) %>%
  st_voronoi()%>%
  st_cast() 

#Plot it plus the original data
ggplot(new_voronoi) +geom_sf() + geom_sf(data=new_data)

#We can clip this to the convex hull easily, but how do we deal with the tails connecting?
new_voronoi %>%
  st_intersection(st_convex_hull(st_combine(new_data)))%>%
  ggplot() +
  geom_sf()+
  geom_sf(data=new_data)