2019-06-18
Want to use a flag (or any *.png file) as a background of your map? You are in the right post. I am aware that there are some R packages out there that are designed to do this, but I decided to challenge myself and see if I could make it from scratch.
After some research on the web I found an article of Paul Murrell, “Raster Images in R Graphics” (The R Journal, Volume 3/1, June 2011), which was a good starting point. With a little bit more of trial and error, I finally got a solution based in the rasterpackage and some support of the png package.
As usual, I wrapped all this research in a function, this time called stdh_png2map, that is stored in my functions folder. This function is designed to work with any *.png file, so you can use logos, personal pictures, etc. with it.
In this post I am going to plot a map of Spain with its autonomous communities (plus 2 autonomous cities), that is the first-level administrative division of the country. Wikipedia shows an initial map identifying also the flag of each territory.
The first step is to choose a good map source for my plot. I really like the geodata available in Eurostat, with information of countries in several formats (*.geojson,*.shp, etc. ), projections (EPSG 4326, 3035 and 3857) and resolutions (01M, 03M, 10M, 20M and 60M), accesible via API.
The map I wanted to work with corresponds to NUTS level 2 for Spain, that matches the autonomous regions of Spain. This shape is also available in Eurostat via API.
Note: NUTS (Nomenclature of Territorial Units for Statistics) is the european model for geopolitical subdivisions.
SPAIN <- st_read(
"https://ec.europa.eu/eurostat/cache/GISCO/distribution/v2/nuts/geojson/NUTS_RG_03M_2016_3857_LEVL_2.geojson",
stringsAsFactors = FALSE
) %>%
subset(CNTR_CODE == "ES")
WORLD <- st_read(
"https://ec.europa.eu/eurostat/cache/GISCO/distribution/v2/countries/geojson/CNTR_RG_03M_2016_3857.geojson",
stringsAsFactors = FALSE
)
# Plot
par(mar = c(0, 0, 0, 0))
plot(st_geometry(SPAIN),
col = NA,
border = NA,
bg = "#C6ECFF"
)
plot(st_geometry(WORLD),
col = "#E0E0E0",
bg = "#C6ECFF",
add = T
)
plot(st_geometry(SPAIN), col = "#FEFEE9", add = T)
layoutLayer(
title = "",
frame = FALSE,
scale = 500,
sources = "© EuroGeographics for the administrative boundaries",
author = "dieghernan, 2019",
)
First thing to note is that the Canary Islands are still in the same place ;) , far away of Spain’s mainland. Just for plotting purposes I am going to place an inset-like map by moving them closer and drawing a box around.
# Move Canary Islands
CAN <- SPAIN %>% subset(NUTS_ID == "ES70")
CANNEW <- st_sf(st_drop_geometry(CAN),
geometry = st_geometry(CAN) + c(550000, 920000)
)
st_crs(CANNEW) <- st_crs(SPAIN)
# New version of Spain
SPAINV2 <- rbind(
SPAIN %>%
subset(NUTS_ID != "ES70"),
CANNEW
)
# BBoxCan
bboxcan <- CANNEW %>% st_bbox()
coords <- cbind(bboxcan[1], bboxcan[4] + 50000)
coords <- rbind(coords, cbind((bboxcan[3] - 20000), bboxcan[4] + 50000))
coords <- rbind(coords, cbind((bboxcan[3] + 40000), bboxcan[4]))
coords <- rbind(coords, cbind(bboxcan[3] + 40000, bboxcan[2]))
bboxcan <- st_linestring(coords) %>%
st_sfc(crs = st_crs(CANNEW))
# Plot
par(mar = c(0, 0, 0, 0))
plot(st_geometry(SPAINV2),
col = NA,
border = NA,
bg = "#C6ECFF"
)
plot(st_geometry(WORLD),
col = "#E0E0E0",
bg = "#C6ECFF",
add = T
)
plot(st_geometry(SPAINV2), col = "#FEFEE9", add = T)
plot(st_geometry(bboxcan), add = T)
layoutLayer(
title = "",
frame = FALSE,
sources = "© EuroGeographics for the administrative boundaries",
author = "dieghernan, 2019",
)
Now we have it! A nice map of Spain with a layout based on the Wikipedia convention for location maps.
As a first example, I chose Asturias to build my code. So the goal here is to create a RasterBrick from the desired *.png file, add the necessary geographical information and use the shape of Asturias to crop the flag. It seems easy, but I needed to sweat to get the following code ready and working:
# 1.Shape---
shp <- SPAINV2 %>% subset(id == "ES12") # Asturias
# 2.Get flag---
url <- "https://upload.wikimedia.org/wikipedia/commons/thumb/3/3e/Flag_of_Asturias.svg/800px-Flag_of_Asturias.svg.png"
dirfile <- paste(tempdir(), "flag.png", sep = "/")
curl_download(url, dirfile)
# 3.Raster---
flag <- brick(readPNG(dirfile) * 255)
# Geotagging the raster
# Adding proj4string
projection(flag) <- CRS(st_crs(shp)[["proj4string"]])
# Now cover with the flag the whole extent of the shape
ratioflag <- dim(flag)[2] / dim(flag)[1]
# Middle point
extshp <- extent(shp)
w <- (extshp@xmax - extshp@xmin) / 2
h <- (extshp@ymax - extshp@ymin) / 2
w_mp <- extshp@xmin + w
h_mp <- extshp@ymin + h
# Depending of the shape the fitting could be in height or width
if (w > h * ratioflag) {
new_ext <- c(
extshp@xmin,
extshp@xmax,
h_mp - w / ratioflag,
h_mp + w / ratioflag
)
} else {
new_ext <- c(
w_mp - h * ratioflag,
w_mp + h * ratioflag,
extshp@ymin,
extshp@ymax
)
}
extent(flag) <- new_ext
par(mar = c(1, 1, 1, 1), mfrow = c(1, 2))
plotRGB(flag)
plot(st_geometry(shp), add = T, lwd = 1.2)
# 4.Mask---
fig <- mask(flag, shp)
plotRGB(fig, bgalpha = 0)
plot(st_geometry(SPAINV2), add = T, lwd = 2)
A note on this: I decided to assign the coordinates of the flag (the extent property) keeping the original aspect ratio and adjusting it to the width or height of the map. In this sense, it is also assessed whether the map shape is “wide” (w > h * ratioflag) or “tall” to determine which dimension should be adjusted.
I have found some interesting things when creating this code:
*.png right through the raster::brick function to create a RasterBrick. The main problem with that is that for *.png files where one channel (Red, Green or Blue) is non present (i.e.monochromatic flags) the RasterBrick doesn’t plot well. For that reason I preferred to import it with readPNG before passing it to the brick function.plotRGB, use the bgalpha=0 option, otherwise the non-masked part of the raster would be plotted as well.lwd=2 in my example).stdh_png2map functionHere you have, a complete solution on a function! I made a final adjustment, the parameter align, that allows to place the mask in the left, center or right side of the *.png, giving the ability to show specific details of the masked figure. This option is only available for “tall” shapes.
Another additional feature is the possibility to use either local *.png files or add them via url.
stdh_png2map <- function(sf, png, align = "center") {
shp <- sf
if (file.exists(png)) {
flag <- brick(readPNG(png) * 255)
} else {
dirfile <- paste(tempdir(), "flag.png", sep = "/")
curl_download(png, dirfile)
flag <- brick(readPNG(dirfile) * 255)
}
if (!align %in% c("left", "right", "center")) {
stop("align should be 'left','right' or 'center'")
}
# Geotagging the raster
# Adding proj4string
projection(flag) <- CRS(st_crs(shp)[["proj4string"]])
# Now cover with the flag the whole extent of the shape
ratioflag <- dim(flag)[2] / dim(flag)[1]
# Middle point
extshp <- extent(shp)
w <- (extshp@xmax - extshp@xmin) / 2
h <- (extshp@ymax - extshp@ymin) / 2
w_mp <- extshp@xmin + w
h_mp <- extshp@ymin + h
# Depending of the shape the fitting could be in height or width
if (w > h * ratioflag) {
new_ext <- c(
extshp@xmin,
extshp@xmax,
h_mp - w / ratioflag,
h_mp + w / ratioflag
)
} else {
if (align == "center") {
new_ext <- c(
w_mp - h * ratioflag,
w_mp + h * ratioflag,
extshp@ymin,
extshp@ymax
)
} else if (align == "left") {
new_ext <- c(
extshp@xmin,
extshp@xmin + 2 * h * ratioflag,
extshp@ymin,
extshp@ymax
)
} else {
new_ext <- c(
extshp@xmax - 2 * h * ratioflag,
extshp@xmax,
extshp@ymin,
extshp@ymax
)
}
}
extent(flag) <- new_ext
# Mask
fig <- mask(flag, shp)
return(fig)
}Here I present an use case for Andalusia:
# Run it
AND <- stdh_png2map(
SPAINV2 %>% subset(id == "ES61"),
"https://upload.wikimedia.org/wikipedia/commons/thumb/2/20/Flag_of_Andaluc%C3%ADa.svg/800px-Flag_of_Andaluc%C3%ADa.svg.png"
)
# Plot it
par(mar = c(0, 0, 0, 0), mfrow = c(1, 1))
plotRGB(AND)
plot(SPAINV2 %>% subset(id == "ES61") %>% st_geometry(),
add = T, lwd = 2
)
Now, we are ready to have fun with flags. It’s time to make the flag map of the autonomous communities of Spain.
par(mar = c(0, 0, 0, 0))
plot(
SPAINV2 %>%
st_geometry(),
col = NA,
border = NA,
bg = "#C6ECFF"
)
plot(st_geometry(WORLD),
col = "#E0E0E0",
add = T
)
plot(st_geometry(bboxcan),
add = T
)
layoutLayer(
title = "",
frame = FALSE,
sources = "© EuroGeographics for the administrative boundaries",
author = "dieghernan, 2019",
)
# Asturias
plotRGB(
stdh_png2map(
subset(SPAINV2, id == "ES12"),
"https://upload.wikimedia.org/wikipedia/commons/thumb/3/3e/Flag_of_Asturias.svg/800px-Flag_of_Asturias.svg.png"
),
bgalpha = 0,
add = T
)
# Andalucia
plotRGB(AND, add = T, bgalpha = 0)
# Aragon
plotRGB(
stdh_png2map(
subset(SPAINV2, id == "ES24"),
"https://upload.wikimedia.org/wikipedia/commons/thumb/1/18/Flag_of_Aragon.svg/800px-Flag_of_Aragon.svg.png",
align = "left"
),
bgalpha = 0,
add = T
)
# ...more flags
# Go to the source code of this post on GitHub for the full code
plot(st_geometry(SPAINV2),
col = NA,
lwd = 2,
add = T
)
We are done now. If you have suggestion (or you know how to get rid of the annoying white spaces around the shapes) you can leave a comment. As always, if you enjoyed the post you can share it on your preferred social network.