Skip to contents

Create an areal spatial domain for SAR/CAR models

Source: R/areal.R
make_areal_domain.Rd

[Experimental]

Build an areal domain object with adjacency weights and unit labels. The returned object can be supplied to mesh in sdmTMB() in areal SAR/CAR workflows.

Usage

make_areal_domain(
  spatial_domain,
  space_column = NULL,
  id_column = NULL,
  adjacency = c("rook", "queen")
)

Arguments

spatial_domain

A named igraph object or an sf/sfc polygon object.

space_column

Column name in model data that identifies areal unit membership. For sf polygon input, defaults to id_column when id_column is supplied; otherwise defaults to "area".

id_column

Optional column name in an sf polygon spatial_domain containing areal unit IDs. If omitted for sf polygons, stable IDs are generated.

adjacency

Polygon adjacency type for sf polygon input: "rook" for shared edges or "queen" for any touching boundary.

Value

A list with class c("sdmTMBareal", "sdmTMBdomain").

See also

make_areal_grid()

Examples

data(ohio_df)
data(ohio_sf)

domain <- make_areal_domain(ohio_sf, id_column = "county")
domain$n_s
#> [1] 88
domain$unit_names
#>  [1] "Lucas"      "Fulton"     "Geauga"     "Williams"   "Cuyahoga"  
#>  [6] "Ottawa"     "Wood"       "Lorain"     "Sandusky"   "Trumbull"  
#> [11] "Henry"      "Erie"       "Defiance"   "Summit"     "Portage"   
#> [16] "Huron"      "Medina"     "Seneca"     "Paulding"   "Hancock"   
#> [21] "Putnam"     "Mahoning"   "Ashland"    "Crawford"   "Richland"  
#> [26] "Wayne"      "Van Wert"   "Wyandot"    "Stark"      "Columbiana"
#> [31] "Allen"      "Hardin"     "Mercer"     "Carroll"    "Morrow"    
#> [36] "Marion"     "Auglaize"   "Holmes"     "Tuscarawas" "Jefferson" 
#> [41] "Knox"       "Logan"      "Union"      "Shelby"     "Coshocton" 
#> [46] "Delaware"   "Harrison"   "Darke"      "Licking"    "Champaign" 
#> [51] "Guernsey"   "Miami"      "Belmont"    "Muskingum"  "Franklin"  
#> [56] "Madison"    "Clark"      "Noble"      "Fairfield"  "Perry"     
#> [61] "Montgomery" "Preble"     "Monroe"     "Greene"     "Pickaway"  
#> [66] "Morgan"     "Fayette"    "Hocking"    "Washington" "Warren"    
#> [71] "Butler"     "Clinton"    "Athens"     "Ross"       "Vinton"    
#> [76] "Highland"   "Hamilton"   "Clermont"   "Brown"      "Jackson"   
#> [81] "Meigs"      "Pike"       "Adams"      "Gallia"     "Scioto"    
#> [86] "Lawrence"   "Ashtabula"  "Lake"      

# \donttest{
fit <- sdmTMB(
  cases ~ pct_male,
  data = ohio_df,
  mesh = domain,
  spatial_model = "car",
  family = poisson(link = "log"),
  offset = log(ohio_df$pop)
)
# }