ggplot2 and eurostat: row removed due to missing values (geom_text) using geom_sf

Using Eurostat I try to plot some values to the map of the EU, using the following code:

library(tidyverse)
library(eurostat)
library(leaflet)
library(sf)
library(scales)
library(cowplot)
library(ggthemes)

#----------- load dataset -----------
SHP <- get_eurostat_geospatial(resolution = 10, 
                        nuts_level = 0,
                        year = 2021)

#select EU countries only
EU28 <- eu_countries %>% 
  select(geo = code, name)
EU29 <- rbind(EU28, c("NO", "Norge"))  # add Norway since not EU country

SHP_29 <- SHP %>% 
  select(geo = NUTS_ID, geometry) %>% 
  inner_join(EU29, by = "geo") %>% 
  arrange(geo) %>% 
  st_as_sf()


# ------------ create values to map on EU --------------
test = data.table(
  geo=EU29$geo,
  values=c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29)
)
test <- SHP_29 %>% inner_join(test, by="geo") %>% arrange(geo) %>%  st_as_sf()
test$values <- as.numeric(test$values)
testSum= sum(test$values)


#---------- plot ---------- #
test %>% 
  ggplot(aes(fill=test$values))  
  geom_sf()  
  geom_sf_text(aes(label=test$values))  
  scale_fill_distiller(
    palette = 1,
    direction = 1,   # reverse default order
    name = "ID"
  )  
  scale_x_continuous(limits = c(-10, 35))  
  scale_y_continuous(limits = c(35, 70))  
  labs(
    title = "Test",
    subtitle = sprintf("Countries=%s", testSum),
  )  
  theme_void()

This results in a beautiful map. However, the value for Norway (=29) is not plotted on the map. Which is probably due to the second warning message obtained when plotting:

Warning messages:
1: In st_point_on_surface.sfc(sf::st_zm(x)) : st_point_on_surface may not give correct results for longitude/latitude data
2: Removed 1 rows containing missing values (geom_text).

Why does ggplot2 or eurostat skip plotting the value for Norway?

CodePudding user response：

The first warning is saying that you are plotting some points outside the continuous boundary limits you set.

Comment the two scale_#_continuous() lines and you will see #29 is being plotted on top of Svalbard, Norway, near the 80 degree North parallel!

Edit: The way to fix this is to "crop" the data first

crop_factor <- st_bbox(c(xmin = -10, 
                         xmax = 35, 
                         ymax = 35, 
                         ymin = 70),
                       crs = st_crs(test))

test_cropped <- st_crop(test, crop_factor)


test_cropped %>% 
  ggplot(aes(fill=values))  
  geom_sf()  
  geom_sf_text(aes(label=values))

CodePudding user response：

The issue is caused by the fun.geometry argument of

CodePudding user response：

Besides the solution of @m-viking (which I did not try) it is also possible to alter the specific nudge of Norway (or any other country of interest) like this:

test2 <- test %>%
  mutate(my_nudge_x=ifelse(geo=='NO',-15,0) ,
         my_nudge_y=ifelse(geo=='NO',-20,0) ,
         geo = ifelse(geo=='NO','Nudged NO', geo)

test %>% 
  ggplot(aes(fill=test$values))  
  geom_sf()  
  geom_sf_text(aes(label=test$values),
  nudge_x = test2%my_nudge_x,
  nudge_y = test2%my_nudge_y,
  )  
  scale_fill_distiller(
    palette = 1,
    direction = 1,   # reverse default order
    name = "ID"
  )  
  scale_x_continuous(limits = c(-10, 35))  
  scale_y_continuous(limits = c(35, 70))  
  labs(
    title = "Test",
    subtitle = sprintf("Countries=%s", testSum),
  )  
  theme_void()

the coordinates for my_nudge_x and my_nudge_y were found using trial&error.

Thanks to https://community.rstudio.com/t/geom-sf-text-change-position-of-only-one-text-label/73419/4