TITLE: Making inset maps for publication
DATE: 2019-02-20
AUTHOR: John L. Godlee
====================================================================


In scientific publications, often there is a map to describe the
study location. For example in Burton et al. (2013), which is the
first publication I found with an inset map by searching for the key
word “inset map”.

 {IMAGE}


I wanted to be able to make these maps directly in R, rather than
manually exporting each map image and overlaying them in PowerPoint
or Photoshop. The maps I make are in {ggplot2}, so the examples here
all use ggplot2. The example involves making a map of Angola and
adding some points of key cities, then adding an inset map showing
the location of Angola in the world.

First, load some packages:

   library(ggplot2)
   library(rworldmap)

Then create fortified polygons by grabbing country data from
{rworldmap}, using the getMap() function:

   world <- getMap(resolution = "low")
   world_f <- fortify(world)

   angola <- world[world@data$ADMIN == "Angola", ]
   angola_f <- fortify(angola)

Then, make the inset world map, by wrapping the normal ggplot() call
in ggplotGrob():

   world_grob <- ggplotGrob(
       ggplot() +
       geom_polygon(data = world_f, aes(x = long, y = lat, group = group), fill = NA, colour = "black") +
       geom_polygon(data = angola_f, aes(x = long, y = lat), fill = "#F2C81F") +
       theme_void() +
       coord_quickmap() +
       theme(panel.background = element_rect(fill = "white", colour = "black"))
   )

This saves the ggplot object in a format that can be added later as
an annotation_custom().

Then, create some city location data:

   city_location <- data.frame(city = c("Luanda", "Lubango"),
       long = c(13.267, 13.500),
       lat = c(-8.845, -14.918))

Then create a ggplot() object of the country of Angola, with the
city locations added as geom_point():

   angola_ggplot <- ggplot() +
       geom_polygon(data = angola_f, aes(x = long, y = lat),
           fill = "#F2C81F", colour = "black") +
       geom_point(data = city_location, aes(x = long, y = lat), colour = "black", size = 5) +
       geom_label(data = city_location, aes(x = long, y = lat, label = city), hjust = -0.2) +
       theme_classic() +
       coord_quickmap() +
       labs(x = "Longitude", y = "Latitude")

Finally, plot angola_ggplot again, using annotation_custom() to add
the world_grob:

   angola_ggplot +
       annotation_custom(grob = world_grob,
           xmin = 18, xmax = 26,
           ymin = -9, ymax = -4)

annotation_custom() allows you to set the x and y location of the
grob, using the coordinate system on the base plot, which in this
case is decimal latitude and longitude.

 {IMAGE}