28 Is This a Chart or a Map?

In thematic maps you can’t always tell. Try to stretch the limits.

In New Zealand drop down lists of regions are almost always presented from north to south almost like a map. It works because the country is ‘narrow’ from east to west. Picking up on that idea here is a population bar chart ‘map’.

Libraries

Code

library(sf)
library(ggplot2)
library(dplyr)
library(stringr)

Data wrangling

Code

pop <- st_read("data/regional-population.gpkg")
bb <- st_bbox(pop)
hght <- (bb[4] - bb[2]) / nrow(pop)

pop_df <- pop %>% 
  st_centroid() %>%
  st_coordinates() %>%
  as_tibble() %>%
  mutate(Population = pop$pop, Region = pop$region) %>%
  arrange(Y) %>%
  mutate(order = bb[2] + row_number() * hght - hght / 2,
         y = Y, x = X, h = hght, w = Population / 20,
         name = str_remove(Region, " Region"))

The map

Since I’ve organised the data as a data table not a spatial dataset, it makes sense to use ggplot2. The geom_tile function can handle rectangles at given location with specified width and height.

ggplot(pop_df) +
  geom_tile(aes(x = x, y = order, width = w, height = h), 
            fill = "#aaffbb", linewidth = 0) +
  geom_text(aes(x = x, y = order, label = name), hjust = 0.5) +
  coord_equal() +
  labs(title = "Population of New Zealand regions") +
  theme_void() +
  theme(panel.background = element_rect(fill = "#bbccee", linewidth = 0))