In mdsumner/scsf: Common Forms for Simple Features

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

scsf

Convert simple features to a generic common form that is more general and can be used for a wide variety of data structures.

This is work in progress and at a very early stage. More to come.

Example - sf to ggplot2 round trip

library(sf)
## a MULTIPOLYGON layer
nc = st_read(system.file("shape/nc.shp", package="sf"))

The common form is the entity tables, objects, paths, vertices and a link table to allow de-duplication of shared vertices. Currently this de-duplication is done on all coordinate fields, but for most applications it will usually be done only in X-Y.

library(scsf)
nc = st_read(system.file("gpkg/nc.gpkg", package="sf"))


(bmodel <- PATH(nc))

Prove that things work by round-tripping to the PATH model and onto the old fortify approach for ggplot2.

inner_cascade <- function(x) {
  tabnames <- silicate:::join_ramp(x)
  tab <- x[[tabnames[1]]]
  for (ni in tabnames[-1L]) tab <- dplyr::inner_join(tab, x[[ni]])
  tab
}

## this just joins everything back together in one big fortify table
library(dplyr)
tab <- bmodel  %>% inner_cascade()

library(ggplot2)
ggplot(tab) + aes(x = x_, y = y_, group = path) + 
  geom_polygon(aes(fill = AREA)) +  geom_path(lwd = 2, col = "black")

What about polygons with holes and lots of tiny complicated parts.

data("inlandwaters")

iw <- PATH(inlandwaters)

str(iw)

tab <- iw  %>% inner_cascade()

library(ggplot2)
ggplot(tab) + aes(x = x_, y = y_, group = path) + 
  ggpolypath::geom_polypath(aes(fill = Province)) +  geom_path(col = "black") 

ggplot(tab %>% filter(Province == "South Australia")) + aes(x = x_, y = y_, group = path) + 
  ggpolypath::geom_polypath(fill = "dodgerblue") +  geom_path(col = "black") + coord_fixed()