This publish is the fourth of a collection of (most likely) seven on inhabitants points within the Pacific, re-generating the charts I utilized in a keynote speech earlier than the November 2025 assembly of the Pacific Heads of Planning and Statistics in Wellington, New Zealand. The seven items of the puzzle are:
At present’s publish is all about creating this one eye-catching chart, evaluating the variety of folks in a rustic with its diaspora—individuals who ethnically or in any other case establish with the nation however reside abroad:
Because the chart notes, the diaspora numbers are an underestimate as a result of I’ve solely drawn on the New Zealand, Australian and USA censuses, and solely partly for that. For instance, I made a decision the variety of Papua New Guineans residing in New Zealand and USA wasn’t materials and so they haven’t been included. I’m assured this doesn’t change the look of the chart, however clearly if I had been making an attempt to create the absolute best complete estimates I ought to embody these.
It’s a reasonably dramatic story. We will see seven nations with extra folks residing abroad than within the nation itself: Niue, Pitcairn Islands, Prepare dinner Islands, Tokelau, Samoa, Tonga and Marshall Islands. Aside from Marshall Islands, these are all Polynesian. The truth is, Tuvalu is the one Polynesian nation on this assortment that has extra folks residing in-country than abroad (for now—that is prone to change now that Australia has agreed with Tuvalu for a daily annual consumption of individuals by way of lottery).
Notice that the three French territories (New Caledonia, Wallis and Futuna, and French Polynesia), and three American territories (American Samoa, Northern Mariana Islands and Guam) have been excluded from the plot.
For the 4 small nations alongside the underside row of the chart, the distinction is especially vital—an enormous majority of their persons are residing abroad. From my final publish we all know that numerous these are in Auckland. Pitcairn is the one considered one of these 4 that has extra of its diaspora in Australia than New Zealand (there are Pitcairn-identifying folks within the UK too, however not sufficient to make me systematically add within the UK to my information in what was primarily a practical and visible train—see feedback above).
96% of Niueans, 90% of Prepare dinner Islanders and 59% of Marshall Islanders reside abroad.
And for the 4 nations on the high of the chart—considerably bigger and distinctly poorer than most of the others, and three of them Melanesian—we see no vital diaspora, relative to the house inhabitants.
Right here’s the code that creates this bar chart. Notice that the info listed here are typed in by hand (!!) from numerous sources—not one thing I’d usually do, and would by no means advocate apart from these actually “small information” conditions. I’ve checked it as totally as I moderately can, and the model I utilized in my speak that I’m adapting right here was additionally peer reviewed by a piece colleague.
# This script attracts some charts of the diaspora of Pacific island nations and territories.
# It is fairly tough and definitely incomplete. The method was to make use of the census figures
# for resident inhabitants of Pacific islander ancestry at present residing in USA, Australia
# and New Zealand; and evaluate that to populations resideing within the nations themselves.
#
# All kinds of identified limitations which we're ready to reside with for these crude comparisons:
# - totally different reference years (2025 for populations, and census years are 2018, 2020 and 2021)
# - populations residing within the Pacific islands themselves are all ethnicities (e.g. will embody
# Australian-descent folks rsideing in these nations), have not bothered to restrict to only "true" Tongans, Samoans, and so forth
# - not complete e.g. I do know there are some Pitcairn-descended folks in UK however have not included them. And naturally
# there should be many others of those folks in nations aside from Australia, NZ and USA
# - France not included in any respect. No ancestry information in French censuses so this is able to be difficult.
#
# Peter Ellis 2025-11
#---------------------Information prep-------------------------
library(tidyverse)
library(rsdmx)
library(ISOcodes)
# Present populations of PICTs:
pops <- rsdmx::readSDMX("https://stats-sdmx-disseminate.pacificdata.org/relaxation/information/SPC,DF_POP_PROJ,3.0/A..MIDYEARPOPEST._T._T?startPeriod=2025&endPeriod=2025&dimensionAtObservation=AllDimensions") |>
as_tibble() |>
left_join(choose(ISOcodes::ISO_3166_1, Alpha_2, pict = Identify), by = c("GEO_PICT" = "Alpha_2")) |>
choose(pict, pop = obsValue) |>
drop_na()
# out of curiosity what's the whole inhabitants of all PICTs, Austrlaia and NZ collectively? (about 47m):
picts_and_anz <- c(sum(pops$pop), 28.1e6, 5.3e6)
sum(picts_and_anz)
# https://instruments.summaries.stats.govt.nz/ethnic-group/tongan and comparable for 2023 NZ determine
# desk builder for Australian 2021 figures - see `https://uncooked.githubusercontent.com/ellisp/blog-source/refs/heads/grasp/information/totalpercent20bypercent20pacific.csv`
# Wikipedia for US figures, from 2020 census. Seek for e.g. "Palauans in USA wikipedia"
diaspora <- tribble(~pict, ~dest, ~folks,
"Tonga", "New Zealand", 97824,
"Niue", "New Zealand", 34944,
"Tokelau", "New Zealand", 9822,
"Prepare dinner Islands", "New Zealand", 94176,
"Samoa", "New Zealand", 213069,
"Tuvalu", "New Zealand", 6585,
"Fiji", "New Zealand", 25038 + 23808, # contains Fijian Indian
"Papua New Guinea", "Australia", 22668,
"Vanuatu", "Australia", 2380,
"Solomon Islands", "Australia", 2704,
"Kiribati", "Australia", 1263,
"Fiji", "Australia", 48354,
"Nauru", "Australia", 571,
"Prepare dinner Islands", "Australia", 27494,
"Tokelau", "Australia", 2544,
"Tonga", "Australia", 43469,
"Niue", "Australia", 6225,
"Samoa", "Australia", 98022,
"Tuvalu", "Australia", 995,
"Pitcairn", "Australia", 1123,
"Marshall Islands", "USA", 52624, # 47300 if simply 'alone'
"Palau", "USA", 12202,
"Micronesia, Federated States of", "USA", 21596)
# Australia checked
# New Zealand checked
# USA checked
#--------------------------Bar chart------------------------
# information body to examine to get percentages.
pops_with_prop <- pops |>
inner_join(diaspora) |>
mutate(pict = gsub("Federated States of", "Fed. St.", pict)) |>
group_by(pict, pop) |>
summarise(Abroad = sum(folks)) |>
ungroup() |>
mutate(prop = Abroad / (pop + Abroad)) |>
mutate(pict = fct_reorder(pict, prop))
pops_with_prop |>
choose(-prop) |>
rename(`Origin nation` = pop) |>
collect(variable, worth, -pict) |>
ggplot(aes(x = variable, y = worth, fill = variable)) +
geom_col(width = 0.8) +
facet_wrap(~pict, scales = "free_y") +
scale_y_continuous(label = comma) +
scale_fill_manual(values = c("steelblue", "brown")) +
theme(legend.place = "none",
panel.spacing = unit(2, "strains"),
plot.caption = element_text(color = "grey50")) +
labs(x = "", y ="Variety of folks",
title = "Pacific Islander diaspora, organized from lowest proportion abroad to highest",
subtitle = "Diaspora is a decrease sure of full determine as it's primarily based on simply Australia, USA and New Zealand censuses.",
caption = "Supply: PDH.Stat for populations; Australian, USA and New Zealand Censuses for diaspora.")Another (not so good) visualisation
I used the identical information to additionally make this scatter plot:
However I don’t very similar to it. It’s troublesome to interpret, and whereas it has a bit of additional info (which nation the diaspora is in) this doesn’t outweigh the interpretation issues. it most likely shouldn’t have a log scale as we actually wish to add up the numbers; however utilizing a non-transformed scale makes it much more of a visible mess. I’m together with it right here actually only for the document and as an example that the primary attempt at visualising one thing isn’t all the time the most effective (and typically, a humble bar chart finally ends up being what you need). Right here’s the code for the scatter plot:
pops |>
inner_join(diaspora) |>
ggplot(aes(x = pop, y = folks, label = pict, color = dest)) +
geom_abline(slope = 1, intercept = 0, color = "gray") +
geom_point() +
geom_text_repel(dimension = 2.5) +
scale_x_log10(label = comma) +
scale_y_log10(label = comma) +
scale_colour_manual(values = c("blue", "black", "darkred")) +
labs(x = "Individuals residing in origin nation in 2025",
y = "Diaspora in a foreign country, latest census",
color = "Disapora nation",
title = "Pacific Island house inhabitants and diaspora in numerous nations")
