random_partition: Define a random partition of the spatial domain based on a...

In spatialstatisticsupna/bigDM: Scalable Bayesian Disease Mapping Models for High-Dimensional Data

Define a random partition of the spatial domain based on a regular grid

Description

The function takes an object of class SpatialPolygonsDataFrame or sf and defines a random partition of the spatial polygons based on a regular grid over the whole domain using the st_make_grid function of the sf package.

Usage

random_partition(
  carto,
  rows = 3,
  columns = 3,
  min.size = 50,
  max.size = 1000,
  prop.zero = NULL,
  O = NULL
)

Arguments

carto	object of class SpatialPolygonsDataFrame or sf.
rows	integer; number of rows to define the regular grid. Default to 3.
columns	integer; number of columns to define the regular grid. Default to 3.
min.size	numeric; value to fix the minimum number of areas in each spatial partition (if NULL, this step is skipped). Default to 50.
max.size	numeric; value to fix the maximum number of areas in each spatial partition (if NULL, this step is skipped). Default to 600.
prop.zero	numeric; value between 0 and 1 that indicates the maximum proportion of areas with no cases for each spatial partition.
O	character; name of the variable that contains the observed number of disease cases for each areal units. Only required if prop.zero argument is set.

Details

After defining a random partition of the spatial polygons based on a regular grid, the subregions with number of areas smaller than the value given by the min.size are merged to its nearest neighbour. Then, the subregions with number of areas greater than the value given by the max.size argument are divided. Finally, if prop.zero argument is set, the subregions with proportion of areas with zero cases below that threshold are merged to its smallest neighbour.

Value

sf object with the original data and a grouping variable named 'ID.group'

Examples

## Not run: 
library(tmap)
tmap4 <- packageVersion("tmap") >= "3.99"

## Load the Spain colorectal cancer mortality data ##
data(Carto_SpainMUN)

## Random partition based on a 3x3 regular grid (with no size restrictions) ##
carto.r1 <- random_partition(carto=Carto_SpainMUN, rows=3, columns=3,
                             min.size=NULL, max.size=NULL)
table(carto.r1$ID.group)

part1 <- aggregate(carto.r1[,"geometry"], by=list(ID.group=carto.r1$ID.group), head)

if(tmap4){
        tm_shape(carto.r1) +
                tm_polygons(fill="ID.group",
                            fill.scale=tm_scale(values="brewer.set3"),
                            fill.legend=tm_legend(frame=FALSE)) +
                tm_shape(part1) + tm_borders(col="black", lwd=2) +
                tm_title(text="3x3 regular grid (with no size restrictions)")
}else{
        tm_shape(carto.r1) +
                tm_polygons(col="ID.group") +
                tm_shape(part1) + tm_borders(col="black", lwd=2) +
                tm_layout(main.title="3x3 regular grid (with no size restrictions)",
                          main.title.position="center", main.title.size=1,
                          legend.outside=TRUE)
}

## Random partition based on a 6x4 regular grid (with size restrictions) ##
carto.r2 <- random_partition(carto=Carto_SpainMUN, rows=6, columns=4,
                             min.size=50, max.size=600)
table(carto.r2$ID.group)

part2 <- aggregate(carto.r2[,"geometry"], by=list(ID.group=carto.r2$ID.group), head)

if(tmap4){
        tm_shape(carto.r2) +
                tm_polygons(fill="ID.group",
                            fill.scale=tm_scale(values="brewer.set3"),
                            fill.legend=tm_legend(frame=FALSE)) +
                tm_shape(part2) + tm_borders(col="black", lwd=2) +
                tm_title(text="6x4 regular grid (min.size=50, max.size=600)")
}else{
        tm_shape(carto.r2) +
                tm_polygons(col="ID.group") +
                tm_shape(part2) + tm_borders(col="black", lwd=2) +
                tm_layout(main.title="6x4 regular grid (min.size=50, max.size=600)",
                          main.title.position="center", main.title.size=1,
                          legend.outside=TRUE)
}

## Random partition based on a 6x4 regular grid (with size and proportion of zero restrictions) ##
carto.r3 <- random_partition(carto=Carto_SpainMUN, rows=6, columns=4,
                             min.size=50, max.size=600, prop.zero=0.5, O="obs")
table(carto.r3$ID.group)

part3 <- aggregate(carto.r3[,"geometry"], by=list(ID.group=carto.r3$ID.group), head)

if(tmap4){
        tm_shape(carto.r3) +
                tm_polygons(fill="ID.group",
                            fill.scale=tm_scale(values="brewer.set3"),
                            fill.legend=tm_legend(frame=FALSE)) +
                tm_shape(part3) + tm_borders(col="black", lwd=2) +
                tm_title(text="6x4 regular grid (min.size=50, max.size=600, prop.zero=0.5)")
}else{
        tm_shape(carto.r3) +
                tm_polygons(col="ID.group") +
                tm_shape(part3) + tm_borders(col="black", lwd=2) +
                tm_layout(main.title="6x4 regular grid (min.size=50, max.size=600, prop.zero=0.5)",
                          main.title.position="center", main.title.size=1,
                          legend.outside=TRUE)
}

## End(Not run)

spatialstatisticsupna/bigDM documentation built on Aug. 22, 2024, 12:35 p.m.