ggmap does not seem to work with ggplot2::geom_sf

[yeedle]

This might be an issue with geom_sf or an issue with how ggmap sets up the dataframe for ggmap or something else. I'm not sure, so I'm submitting it here and to ggplot2

library(sf)
library(ggmap)

nc_map <- get_map(c(-81, 35), zoom = 6) 
#> Map from URL : http://maps.googleapis.com/maps/api/staticmap?center=35,-81&zoom=6&size=640x640&scale=2&maptype=terrain&language=en-EN&sensor=false
ggmap(nc_map)

nc <- st_read(system.file("shape/nc.shp", package="sf"))
ggplot(nc) + geom_sf()

ggmap(nc_map) + geom_sf(data = nc)
#> Error in FUN(X[[i]], ...): object 'lon' not found

[lepennec]

I've made an attempt to solve the issue with pull request #162.

Note that for the nc shape, I have a slight offset between the shape and the map... so that some issues may remain.

[scottmmjackson]

I think we've got a decision to make here. @lepennec's solution seems to work, but I'm not sure what side effects it might have making the mapping specific to the geom_blank call.

@hadley proposed an alternative solution in the ggplot2 dupe issue. We may want to look at that as well.

[gregleleu]

Hi,
I'm also trying to use geom_sf with ggmap and can't get the raster and layer to superpose correctly.
I thought it was a CRS issue but they seem coherent and nothing I tried worked.
@lepennec, is that the offset you mentionned?

Code and output below if it can help.

# UK local authorities boudary from UK geoportal (using KML for easy url version, it does the same with a shapefile)
uk_lads <- st_read("https://opendata.arcgis.com/datasets/fab4feab211c4899b602ecfbfbc420a3_3.kml")
st_crs(uk_lads)
# Coordinate Reference System:
#   EPSG: 4326 
#   proj4string: "+proj=longlat +datum=WGS84 +no_defs"
test_map_uk <- ggmap::get_map(location = unname(st_bbox(uk_lads)), source = "stamen")
ggmap::ggmap(test_map_uk) +
  geom_sf(data = uk_lads, inherit.aes = FALSE)

"Regular" plotting works well:

plot(st_transform(uk_lads,3857)[1], bgMap = test_map_uk)

[adrfantini]

@gregleleu I also had this problem and never found the solution, unfortunately. Other map types work, just google has this problem.

[ateucher]

After struggling with this for a while, I found a way: The bounding box of the ggmap object is in WGS84 (EPSG:4326), but the actual raster is in EPSG:3857. You have to hack the bounding box of the ggmap object to be in the same CRS as the underlying data.

library(sf)
#> Linking to GEOS 3.6.2, GDAL 2.3.0, proj.4 5.1.0
library(ggmap)
#> Loading required package: ggplot2
#> Google Maps API Terms of Service: http://developers.google.com/maps/terms.
#> Please cite ggmap if you use it: see citation("ggmap") for details.
library(ggplot2)

uk_lads <- st_read("https://opendata.arcgis.com/datasets/fab4feab211c4899b602ecfbfbc420a3_3.kml")

# Convert to 3857
uk_lads_3857 <- st_transform(uk_lads, 3857)

test_map_uk <- ggmap::get_map(location = unname(st_bbox(uk_lads)), source = "stamen")

# overwrite the bbox of the ggmap object with that from the uk map that has
# been transformed to 3857
attr(test_map_uk, "bb")$ll.lat <- st_bbox(uk_lads_3857)["ymin"]
attr(test_map_uk, "bb")$ll.lon <- st_bbox(uk_lads_3857)["xmin"]
attr(test_map_uk, "bb")$ur.lat <- st_bbox(uk_lads_3857)["ymax"]
attr(test_map_uk, "bb")$ur.lon <- st_bbox(uk_lads_3857)["xmax"]

ggmap::ggmap(test_map_uk) +
  coord_sf(crs = st_crs(3857)) + # force it to be 3857
  geom_sf(data = uk_lads_3857, inherit.aes = FALSE)
#> Coordinate system already present. Adding new coordinate system, which will replace the existing one.
#> Coordinate system already present. Adding new coordinate system, which will replace the existing one.

Created on 2018-06-13 by the reprex package (v0.2.0).

[ateucher]

Edit: I posted a more general solution on StackOverflow which gets the bounding box from the ggmap object itself rather than from another source.

library(ggplot2)
library(ggmap)
library(sf)

uk_lads <- st_read("https://opendata.arcgis.com/datasets/fab4feab211c4899b602ecfbfbc420a3_3.kml")

# Convert to 3857
uk_lads_3857 <- st_transform(uk_lads, 3857)

test_map_uk <- ggmap::get_map(location = unname(st_bbox(uk_lads)), source = "stamen")

# Define a function to fix the bbox to be in EPSG:3857
ggmap_bbox <- function(map) {
  if (!inherits(map, "ggmap")) stop("map must be a ggmap object")
  # Extract the bounding box (in lat/lon) from the ggmap to a numeric vector, 
  # and set the names to what sf::st_bbox expects:
  map_bbox <- setNames(unlist(attr(map, "bb")), 
                       c("ymin", "xmin", "ymax", "xmax"))
  
  # Coonvert the bbox to an sf polygon, transform it to 3857, 
  # and convert back to a bbox (convoluted, but it works)
  bbox_3857 <- st_bbox(st_transform(st_as_sfc(st_bbox(map_bbox, crs = 4326)), 3857))
  
  # Overwrite the bbox of the ggmap object with the transformed coordinates 
  attr(map, "bb")$ll.lat <- bbox_3857["ymin"]
  attr(map, "bb")$ll.lon <- bbox_3857["xmin"]
  attr(map, "bb")$ur.lat <- bbox_3857["ymax"]
  attr(map, "bb")$ur.lon <- bbox_3857["xmax"]
  map
}

# Use the function:
test_map_uk <- ggmap_bbox(test_map_uk)

ggmap(test_map_uk) + 
  coord_sf(crs = st_crs(3857)) + # force the ggplot2 map to be in 3857
  geom_sf(data = uk_lads_3857, inherit.aes = FALSE)

[thiesben]

I honestly think this should be implemented somewhere as long as there is no other solution available.

[gregleleu]

I submitted a pull request (also including base maps from carto) but it adds an sf dependency which fails at build time because some libraries are not on the CI/CD system

[JosiahParry]

Any movement here? :o

[admahood]

I honestly have been trying to figure out this hack for like an hour and I still can't get it to work

[talhazafarjutt]

coord_sf(crs = st_crs(3857)) + # force the ggplot2 map to be in 3857
  geom_sf(data = uk_lads_3857, inherit.aes = FALSE)

what will be the value of crs if we use map of singapore