The census bureau doesn\'t provide a nationwide shapefile of public use microdata areas (the smallest geography available on the American Community Survey). I tried combining t
Here's another approach, which includes a short cut for obtaining the FTP directory listing. As @Pop mentioned, the key is to ensure that the IDs are all unique.
library(RCurl)
library(rgdal)
# get the directory listing
u <- 'ftp://ftp2.census.gov/geo/tiger/TIGER2014/PUMA/'
f <- paste0(u, strsplit(getURL(u, ftp.use.epsv = FALSE, ftplistonly = TRUE),
'\\s+')[[1]])
# download and extract to tempdir/shps
invisible(sapply(f, function(x) {
path <- file.path(tempdir(), basename(x))
download.file(x, destfile=path, mode = 'wb')
unzip(path, exdir=file.path(tempdir(), 'shps'))
}))
# read in all shps, and prepend shapefile name to IDs
shps <- lapply(sub('\\.zip', '', basename(f)), function(x) {
shp <- readOGR(file.path(tempdir(), 'shps'), x)
shp <- spChFIDs(shp, paste0(x, '_', sapply(slot(shp, "polygons"), slot, "ID")))
shp
})
# rbind to a single object
shp <- do.call(rbind, as.list(shps))
# plot (note: clipping to contiguous states for display purposes)
plot(shp, axes=T, xlim=c(-130, -60), ylim=c(20, 50), las=1)
# write out to wd/USA.shp
writeOGR(shp, '.', 'USA', 'ESRI Shapefile')
