See Georgios Karamanis's example of doing this by converting the map into a raster object
library(tidyverse)
library(sf)
library(rnaturalearth)
# Map data
world_sf <- ne_countries(returnclass = "sf")
# Some cities and places
places <- tribble(
~place, ~lat, ~long,
"Australia", -25.2744, 133.7751,
"Greenland", 76.7069, -40.6043,
"Congo-Amazon\nrainforest", -8, 45,
"Quebec", 55.9399, -65.5491,
"Andes", -10.8895, -68.8458,
"Hawaii", 24.8968, -158.5828,
"Rocky Mountains", 48.5501, -110.9027,
"Titanic Wreck", 40.7269, -49.9483,
"Franklin's\nVery Lost\nExpedition", 69, -97,
"St. Helena", -15.9583, -5.7020,
"Palk-Panama Canal", 10, -79
) %>%
st_as_sf(coords = c("long", "lat"), crs = st_crs("EPSG:4326")) # WGS 84
# Basic map
ggplot() +
geom_sf(data = world_sf) +
geom_sf_text(
data = places, aes(label = place),
lineheight = 0.9
) +
coord_sf(crs = st_crs("ESRI:54030")) + # Robinson
theme_void()
# -------------------------------------------------------------------------
# Step 1: Make all the longitude points in the map use the absolute value
# -------------------------------------------------------------------------
# The geometry column contains MULTIPOLYGON shapes that are really just matrices
# of geographic points within deeply nested lists, like this:
example <- list(list(list(matrix(c(-1, 2, 3, 4), ncol = 2)),
list(matrix(c(5, -3, 7, 8), ncol = 2))),
list(list(matrix(c(-9, -4, 11, 12), ncol = 2))))
example
#> [[1]]
#> [[1]][[1]]
#> [[1]][[1]][[1]]
#> [,1] [,2]
#> [1,] -1 3
#> [2,] 2 4
#>
#>
#> [[1]][[2]]
#> [[1]][[2]][[1]]
#> [,1] [,2]
#> [1,] 5 7
#> [2,] -3 8
#>
#>
#>
#> [[2]]
#> [[2]][[1]]
#> [[2]][[1]][[1]]
#> [,1] [,2]
#> [1,] -9 11
#> [2,] -4 12
# We need to get the absolute value of the first column in each of these deeply
# nested lists
#
# I tried doing this with purrr functions and with a bunch of nested lapply()s,
# but it was still a huge mess. BUT base R has rapply(), which recursively works
# through nested lists and does stuff to each element
#
# So here, we go through all the nested elements of this gross list. If the
# nested object is a matrix, we'll take the absolute value of the first column
rapply(example, function(x) {
if (is.matrix(x)) {
x[, 1] <- abs(x[, 1])
}
x
}, how = "replace")
#> [[1]]
#> [[1]][[1]]
#> [[1]][[1]][[1]]
#> [,1] [,2]
#> [1,] 1 3
#> [2,] 2 4
#>
#>
#> [[1]][[2]]
#> [[1]][[2]][[1]]
#> [,1] [,2]
#> [1,] 5 7
#> [2,] 3 8
#>
#>
#>
#> [[2]]
#> [[2]][[1]]
#> [[2]][[1]][[1]]
#> [,1] [,2]
#> [1,] 9 11
#> [2,] 4 12
# This also works with the geometry column. The first column in the MULTIPOLYGON
# matrix is longitude, so we'll get its absolute value
world_sf_oops <- world_sf
world_sf_oops$geometry <- rapply(st_geometry(world_sf), function(x) {
if (is.matrix(x)) {
x[, 1] <- abs(x[, 1])
}
x
}, how = "replace")
# yesssss
# Though the original xkcd map does better with West Africa and Western Europe,
# and with Antarctica—those shapes go all the way to the edge of the map, but
# not here. Alas.
ggplot() +
geom_sf(data = world_sf_oops) +
theme_void()
# It would be neat to use st_union() to combine all these overlapping countries
# into one big blob, but I can't figure out how :( There are errors because of
# the weird country overlaps, even when using `st_make_valid()`
world_sf_blob <- world_sf %>% st_make_valid() %>% st_union()
#> Error in wk_handle.wk_wkb(wkb, s2_geography_writer(oriented = oriented, : Loop 0 is not valid: Edge 0 crosses edge 78
# So we can fake it by not plotting any boundaries
ggplot() +
geom_sf(data = world_sf_oops, linewidth = 0) +
theme_void()
# ------------------------------------------------------------------------------------
# Step 2: Make all the longitude points in the list of places use the absolute value
# ------------------------------------------------------------------------------------
# The geometry column in `places` contains POINT objects that are just
# two-element vectors. There's no weird deep nesting, so we can just use
# purrr::map() (though we need to wrap it in some sf-related functions to
# extract the geometry column and reproject it to WGS 84 when we're done)
places_oops <- places %>%
mutate(geometry = st_sfc(map(st_geometry(.), ~{
.x[1] <- abs(.x[1])
.x
}))) %>%
st_set_crs(st_crs("EPSG:4326"))
ggplot() +
geom_sf(data = world_sf_oops, linewidth = 0, fill = "white") +
geom_sf_text(data = places_oops, aes(label = place), lineheight = 0.75) +
theme_void() +
theme(panel.background = element_rect(fill = "grey80"))
# The cool thing about leaving all this data as sf objects instead of raster
# data is that we can still do all sf-related stuff, like using different
# projections
ggplot() +
geom_sf(data = world_sf_oops, linewidth = 0, fill = "white") +
geom_sf_text(data = places_oops, aes(label = place), lineheight = 0.75) +
coord_sf(crs = st_crs("ESRI:54030")) + # Robinson
theme_void() +
theme(panel.background = element_rect(fill = "grey80"))
ggplot() +
geom_sf(data = world_sf_oops, linewidth = 0, fill = "white") +
geom_sf_text(data = places_oops, aes(label = place), lineheight = 0.75) +
coord_sf(crs = st_crs("ESRI:54008")) + # Sinusoidal
theme_void() +
theme(panel.background = element_rect(fill = "grey80"))
Created on 2023-08-01 with reprex v2.0.2
UPDATE thanks to Georgios! If you run
sf_use_s2(FALSE)to disable spherical projections, you can merge all the countries into one big blob:Created on 2023-08-01 with reprex v2.0.2