Recently I have been using the sf and tmap packages a lot and I thought I would come back to finish a draft post I started a while ago.

As part of a project a while ago, I needed to create a square grid across Scotland and produce centroids for each. We start by loading the packages we’ll need and the shapefile.

library(sf)
library(tmap)
scot <- st_read("scot_shapefile/scot.shp")

We can see what this looks like with a quick plot.

tm_base <- tm_shape(scot) + 
  tm_polygons()
tm_base

Next we define the size of the grid (in metres) and use st_make_grid() to create the grid based on the bounding box.

size <- 50000 # 50km squares to speed things up

# create grid
scot_grid <- st_make_grid(scot, square = TRUE, cellsize = c(size, size)) %>%
  st_sf()

We can plot this to see what it looks like.

tm_base +
  tm_shape(scot_grid) + 
  tm_borders() # just the borders of the grid cells

Now we have a choice. We can calculate all the centroids and then keep only those that intersect with Scotland or find the intersection first. Here we use st_filter() to subset the grid to just those squares that intersect with Scotland.

scot_only_grid <- st_filter(scot_grid, scot) 
tm_base +
  tm_shape(scot_only_grid) + 
  tm_borders() # just the borders of the grid cells

Now we can calculate the centroids and plot everything to demonstrate.

# calculate centroids for each square in the grid
grid_points <- st_centroid(scot_only_grid)

# with tmap
tm_base +
  tm_shape(scot_only_grid) + 
  tm_borders() + 
  tm_shape(grid_points) + 
  tm_dots(size = 0.25) # making them a little more noticeable