Neighboring Zipcodes from State Polygons - r

Question

First time posting on SO
I have a shapefile that has the geometries for each Zipcode along with state name. I want to figure out which zipcodes lie on the state borders.
The way I figured to achieve this is by combining all zipcodes for each state and leading to the geometry for a state and then finding the neighboring zipcodes for each state.
I combined the zipcodes into states using:
state_shape <- shapefile %>% group_by(State) %>% summarise(geometry = sf::st_union(geometry))
But then when I try to find the neighboring zipcodes using poly2nb
state_nb <- poly2nb(st_geometry(state_shape))
It gives me an Error:
Error in poly2nb(st_geometry(state_shape)) : Polygon geometries required
I understand to find the border zipcodes I will have to pass the zipcode geometries in poly2nb, but the error persists.
Any help will be highly appreciated, also any other approaches to this problem are more than welcome.

Answer 1

Consider this example, built on the widely available North Carolina shapefile that is distributed with {sf} package.
What the example does is:
creates a border line of North Carolina by first dissolving the counties, and then casting the resulting multipolygon to a multilinestring
runs sf::st_touches() on the counties and borderline with sparse set to false; the result is a logical vector that can be used to subset the original shapefile (filtering out the counties that share a border with the NC border)
presents the results in a graphical format, using {ggplot2}; the bordering counties are blue and the rest just blank for context
library(sf)
library(dplyr)
library(ggplot2)
# all NC counties (from shapefile distributed with {sf})
shape <- st_read(system.file("shape/nc.shp", package="sf"))
# border, via dplyr::summarise() & cast as a linestring
border <- shape %>%
summarise() %>%
st_cast("MULTILINESTRING")
# logical vector of length nrow(shape)
neighbours <- sf::st_touches(shape,
border,
sparse = F)
# report results
ggplot() +
geom_sf(data = shape[neighbours, ], fill = "blue") + # border counties
geom_sf(data = shape, fill = NA, color = "grey45") # all counties for context