Mapping Australia’s 2023 Indigenous Voice Referendum, Part I

On October 14, 2023, Australian voters decisively rejected a proposed constitutional amendment that would have recognized the country’s indigenous population by creating a federal advisory body to represent the views of Aboriginal and Torres Strait Islander communities. The 60 percent “no” vote shocked many Australian, as early in the year polling indicated that almost two-thirds of Australians supported the measure. The referendum’s defeat has resulted in a good deal of soul-searching, as well as accusations of misleading campaigning and outright prevarication by those opposed to the measure.

Although these are important issues, the next few GeoCurrents posts will consider instead the lessons that might be learned about Australian electoral geography from this hotly contested referendum. We will also consider what the vote pattern can tell us about the changing nature of Australia’s main political parties and the voting-blocks that support them. As we shall see, although the governing Labour Party strongly supported the measure, many heavily Labour-voting electoral divisions rejected it by considerable margin. At the same time, several important electoral divisions that have historically been strong supporters of the center-right Liberal Party, which opposed the measure, voted in its favor.

Today’s initial post, however, takes on a much simpler and more familiar issue: the tendency for electoral maps to exaggerate support for conservative parties and positions by giving undue visual weight to low-density, rural areas. Consider, for example, Wikipedia’s map of the election results (below). The is a poor example of the cartographer’s craft, as it lacks a key or any other form of explanation. But one can easily infer that darker shades of red indicate a strong “no” vote, whereas the small green area – Canberra, or the Australian Capital Territory – voted “yes.” The overall impression conveyed by this map is that the election was a landslide, with almost all constituencies voting against the measure.

A vastly better map was posted on Reddit’s MapPorn forum – as is so often the case. Unfortunately, however, this map misrepresents the vote in the Northern Territory, where 60.3 percent of voters opposed the measure. But by expanding the few relatively densely populated parts of the country, the map accurately shows widespread support for the referendum in metropolitan areas, where the most Australians live. Melbourne in particular is revealed as a stronghold for the “yes” vote. But the demographic imbalances in Australia are so extreme that this map still does not do justice to the actual vote. As the next set of maps illustrates, Australia’s two largest metropolitan areas, Sydney and Melbourne, together have more than twice the population of the entire western two-thirds of the country. In this vast region, only two electoral divisions, both in Perth, voted “yes,” whereas 17 did so in greater Melbourne and Sydney.

To adequately capture the demographic geography of this election, a cartogram* must be used instead. I was only able to find one example, a mosaic cartogram from the Australian Broadcasting Corporation (ABC) in which of the country’s 151 federal electoral divisions are depicted as hexagons of equal size (although these divisions vary slightly in population, they all have roughly the same number of voters). The cartogram on the right (below) gives a particularly good visual representation of the demographic patterns found in this election.

The problem with cartograms, however, is that of spatial representation. All cartograms distort size and shape, but the issue is often pronounced in places with extremely uneven distributions of population, such as Australia. In the ABC mosaic cartogram posted above, the “geobody” of the country becomes unrecognizable. As the next map shows, it also misrepresents spatial positions. The electoral division of Griffith, for example, appears to be located in central Queensland, but it is actually situated in the state’s far southeastern corner.

All such problems, however, are intrinsic to electoral mapping. My preferred response is to use a variety of maps, made at different scales, and compare them. The next few GeoCurrents posts will do exactly that for Australia’s Indigenous Voice referendum. For now, however, I would like to note that the ABC article that posted the cartograms discussed above also includes several excellent graphs of the election results. Two of these are posted below. Together, they show that the “no” vote was especially pronounced in remote areas with relatively low rates of educational attainment. These correlations, and more, will be explored in greater detail in a set of maps focused on the Sydney metropolitan area that will be posted on this website next soon.


*As defined by Wikipedia: A cartogram (also called a value-area map or an anamorphic map, the latter common among German-speakers) is a thematic map of a set of features (countries, provinces, etc.), in which their geographic size is altered to be directly proportional to a selected ratio-level variable, such as travel time, population or GNP. Geographic space itself is thus warped, sometimes extremely, in order to visualize the distribution of the variable. It is one of the most abstract types of map; in fact, some forms may more properly be called diagrams.

Mapping Australia’s 2023 Indigenous Voice Referendum, Part I Read More »

Surprising Findings in a Study of Post-COVID Urban Recovery Rates in the United States and Canada

I recently came across a brief report by the University of Toronto’s School of Cities on the recovery of urban cores in the U.S. and Canada since the COVID-19 pandemic. The study’s methodology is intriguing:

The recovery metrics on these charts are based on a sample of mobile phone data. The recovery metrics on the charts and maps are computed by counting the number of unique visitors in a city’s downtown area in the specified time period (standardized by region), and then dividing it by the standardized number of unique visitors during the equivalent time period in 2019. Specifically, the rankings below compare the period from the beginning of March to mid-June in 2023 relative to the same period in 2019. A recovery metric greater than 100% means that for the selected inputs, the mobile device activity increased relative to the 2019 comparison period. A value less than 100% means the opposite, that the city’s downtown has not recovered to pre-COVID activity levels.

As the results were given in tabular form, I thought that it would be useful to map them to more easily see if there are any distinct regional patterns or anomalies. The resulting map, posted below, has some expected features. Recovery has generally been faster in low-density sunbelt cities, with only Las Vegas showing an increase in downtown activity since 2019.

But there were also some unexpected findings. Columbus, Ohio, for example, has many characteristics of a sunbelt city, despite its cloudy winters, yet it has one of the worst downtown recovery rates. Minneapolis and Seattle also have unexpectedly low rankings. To understand what is going on in these cities one would have to examine exactly how “downtown” is defined in each case. In Columbus, for example, the old central business district has been declining over the past few decades, but a new vibrant urban core has emerged nearby, in a neighborhood dubbed “Short North.” I doubt that it was included in area deemed downtown Columbus.

A few other interesting findings deserve comment. It seems that Canadian downtowns have recovered more quickly on average than those of the northern United States. Is this because they tended to be more alive to begin with? The relatively quick recovery of Oakland, California that is indicated by the study makes little sense. From what I have read of Oakland, and from what I have seen in a few quick visits, the city’s downtown is in a desperate situation, with closing businesses and surging crime. In late September, 2023, Oakland saw an unprecedented strike of business owners. As reported by a local news source:

It’s not business as usual in downtown Oakland on Tuesday morning as store and restaurant owners go on strike over rising crime.

Business owners say the goal of this strike is to send a larger message to City Hall. They want better protection and support so they can safely operate their businesses and make a living.

Many of the participating businesses gathered in front of Le Cheval for a news conference on Tuesday to voice their concerns. The restaurant is closing at the end of the month because of the crime and slow sales post-pandemic.

Participating merchants say, just like Le Cheval, they’re losing customers and foot traffic because of car break-ins, carjackings, robberies and assaults.

In conclusion, I can only fall back of the most tiresome of all academic clichés: more research is needed.


Surprising Findings in a Study of Post-COVID Urban Recovery Rates in the United States and Canada Read More »

Journalistic Hyperbole and the Electoral Geography of Poland

Although The Economist magazine is to be commended for analyzing and mapping the role of Poland’s old imperial divisions in its current electoral geography, it succumbs to unnecessary and misleading simplification and exaggeration, as is so often the case when journalists take on intricate geographical issues. Consider, for example, the following assertion:

More remarkable, however, is that rather than following a gradual gradient from east to west, modern Poles’ political loyalties remain firm right up to the edge of a historical line that cuts down the middle of the country.

This vivid statement is incorrect, as is demonstrated by the maps included in the two previous GeoCurrents posts. A detail from one of these maps (below) drives the point home. As can be seen, there is only a vaguely discernable difference in “Poles’ political loyalties” as one cross the old boundary between the German and Russian empires between Łódź and Wroclaw. More significant is the fact that two of the five constituencies that gave the lowest percentage of their votes to United Right, Łódź and Warsaw, and two of the four that give the highest percentage to Civic Coalition (again, Łódź and Warsaw), are located to the east of the dividing line.

The Economist article in question also makes a more interesting observation:

From the air, the former Habsburg and Russian territories look like a patchwork mosaic of small farming plots, whereas the west is divided into sprawling fields designed to facilitate mechanised agriculture.

This assertion is certainly intriguing, but is it true? I tested it by trying to find the old border on satellite images found in Apple Maps. I did so with a modest level of success. Although I could not precisely trace out the divide, I did find a significant general difference in the divisions of agricultural land found in the two sectors. Consider, for example, an image (below) of the Kalisz area, located astride the old boundary. The fields on the left side of this image are larger than those on the right, although not to the extent that The Economist suggests. They are also more irregularly shaped.

If anything, the term “mosaic” fits more closely with the field patterns found in the old German sector, as the various pieces of a mosaic are typically of different size and shape. The two images posted below reveal a more mosaic-like agricultural pattern in the west, and a greater degree of field uniformity in the former Russian zone. What the second image shows is not a “mosaic” composed of disparate pieces, but a rather a patchwork composed of many elongated rectangular pieces of similar size. Such an arrangement reflects the pre-modern farming system that used long, narrow fields to facilitate plowing, as it was difficult to turn plow-teams of oxen or horses around. But although the redivision of the land that occurred in the west facilitated mechanization, it was not a necessary precondition for it.

Journalistic Hyperbole and the Electoral Geography of Poland Read More »

Mapping the Results of Poland’s 2023 Elections by Political Parties and Coalitions

Poland, like many other countries with parliamentary systems of government, has many active political parties, most of which belong to broader coalitions. Five of these coalitions, one on the left and center-left, two in the center, and two on the right, received enough votes to win seats in Poland’s Sejm, its powerful lower house of Parliament. Two additional stand-alone parties together received 3.5 percent of the vote, which was not enough to gain representation in the Sejm. One of these is a hard-right populist party and the other is described by Wikipedia as being on the center-left, although some of its positions are more centrist in orientation.

This post seeks to uncover some subtle aspects of Poland’s electoral geography by mapping the vote-share of each of these seven parties and coalitions in the 2023 Sejm election. For comparative purposes, all of them are mapped in the same color scheme and with the same categories of vote percentage. As a result of their low vote counts, the five secondary parties and coalitions are poorly represented on these maps. I have therefore re-mapped them on their own terms, using different color schemes. On these maps, low votes counts for a given party can still be mapped with dark shades, indicating relative success in that area.

Let us begin with the right-wing-populist United Right coalition, led by the Law and Justice (PiS) Party. Although it suffered a sharp rebuke in this election, United Right still received more votes (35.4 percent) than any of its rivals. United Right was the only organization to receive majority support in any electoral district. As can be seen on the first map below, its support was concentrated in the southern and eastern areas that had been under Russian and Austrian rule before WWI (see the previous post). United Right did poorly in major cities, getting less than 20 percent of the vote in Poznań. This map also reveals, albeit weakly, an electoral gradient in United Right’s main area of support, with its vote share increasing toward Poland’s southeastern borders. As might be expected, its main rival, the centrist Civic Coalition, exhibited an inverted spatial pattern of support, which is revealed in the second map posted below.

The major new player in this election, the centrist Third Way coalition, did not have strong regional patterning. It performed relatively well in some regions that had been ruled by Russia, German, and Austria before the reestablishment of Poland after WWI. Third Way did not do as well in major cities, however, as might be expected for a coalition that has a strong agrarianist bent. The only electoral district in which it failed to gain more than 10 percent of the vote is Katowice III. Although Katowice is not well-known outside of Europe, it is Poland’s largest metropolitan area, with more than three million people living in a group of closely clustered cities. (Metropolitan Warsaw, in contrast, has around two and half million inhabitants, while Krakow and Lodz, the next largest, have only a little more than one million each). Third Way’s vote-share was highest in the Bialystok district in the northeast; as the maps used in the previous post show, it even received a plurality of votes in a few areas in and near the city of Hajnówka, close to the border with Belarus. It is probably not coincidental that Hajnówka has a sizable (over 25 percent) Belarussian ethnic minority, whose members tend to shun the Polish ethnonationalism associated with the country’s right-wing parties.

As the next two maps show, support for Lewica – “The Left” – was strongest in urban areas and weakest in Poland’s more conservative eastern regions. As can be seen on the second map, Katowice III was again an outlier, giving more than 20 percent of its votes to a coalition than failed to crack 15 percent anywhere else in the country. Katowice’s economy has until recently been based on heavy industry, whose workers formed one of The Left’s traditional political bases. But as has been widely noted, that foundation of support has been slipping. As the Wikipedia article on Lewica notes, the coalition has lost votes because its “pro-LGBT rights platform failed to appeal to working class and economically left-leaning Poles, [who] tend to favour a more socially conservative policy (especially as both economically interventionist and social conservative positions were already being provided by the right-wing PiS party).”

The far-right Confederation Liberty and Independence, or simply Confederation, had a low but relatively evenly distributed level of support, receiving between five and 10 percent of the vote in every Polish electoral district. By mapping its results with a finer set of divisions, however, we can see that its support follows the typical pattern of rightwing organizations in Poland, being lower in urban areas and higher in the east. This geographical pattern is not so clear-cut, however, for the newest far-right party, There Is One Poland. Its elevated level of support in Nowy Sącz is difficult to explain. One of Confederation’s component parties is the monarchist and Russophile Confederation of the Polish Crown, which managed to secure two Sejm seats despite receiving less than one percent of the vote.

Finally, we come to the oddly named movement called Nonpartisan Local Government Activists. As its name implies, it is a highly decentralized organization that advocates increased regional and local autonomy. Many of its more concrete proposals have a leftward slant, but others are more conservative. It also has an environmentalist platform. At any rate, Wikipedia describes this organization as follows:

Formerly associated with the liberal wing of Christian democracy, the party advocates for proposals such as free public transport, free lunches for children and abolition of the personal income tax (PIT). The party also advocates for creation of a powerful ecological agency based on the American United States Fish and Wildlife Service, which would protect nature and its resources, prevent pollution environment and combat poaching and illegal hunting. The BS believes that the Polish government became too centralised and became unable to address local concerns; to this end, the party believes that the central government needs MPs who are “local government officials, entrepreneurs and community workers who are not concerned with political lists and agenda.”

As can be seen in the maps below, support for Nonpartisan Local Government Activists varied little across the county. Not surprisingly, it performed best in Legnica, which is located in Lower Silesia, the party’s birthplace. A Silesian autonomy movement has long enjoyed considerable support, buttressed by widespread suspicions about Polish nationalism and feelings of affinity with Germany and Germans. This is a complicated issue, however, that deserves its own post.

The most important aspect of Poland’s electoral geography is weakness of left-wing political parties in all areas of the country. The Polish Green party, for example, received less than one-third of one percent of the vote in the 2023 election, although, as part of the Civic Coalition, it did gain three seats in the Sejm.

Mapping the Results of Poland’s 2023 Elections by Political Parties and Coalitions Read More »

Former Imperial Boundaries and Population Density in Poland’s 2023 Election

Poland’s October 2023 election saw a sharp rebuke to the country’s illiberal, governing right-wing coalition. The United Right (ZP), led by the Law and Justice Party (PiS), saw its vote share* drop from 44.6 percent in 2019 to 35.4 percent, undermining its ability to form a new government. But this election was not a victory of the left, but rather of the center, or perhaps even the center-right, depending on how one classifies some of Poland’s political parties. The democratic-socialist Lewica Party (“The Left”) also saw a sharp decline, its vote-share dropping from 12.6 percent to 8.6 percent. In contrast, a sizable gain was realized by the main oppositional group, the centrist Civic Coalition, whose vote share rose from 27.4 to 30.7 percent. The biggest change, however, was the rise of the new Third Way (TD) coalition, which secured 14.4 percent of the vote. Although Third Way is usually regarded as centrist, the Wikipedia classifies it as center-right. Such discrepancies arise from the fact that this coalition’s various factions are ideologically diverse, some being much more centrist than others. But as Third Way overall is pro-EU and favors renewable energy, it is perhaps most accurate regarded, at least in the Polish context, as firmly in the center. It is also noteworthy, however, that Poland’s extreme-right did relatively well in this election, with the anti-EU Confederation for Liberty and Independence taking 7.2 percent of the vote, up from 6.8 in 2019, and the new There Is One Poland (PJJ) gaining another 1.6 percent. The PJJ party, which claims to be the “true right,” grew out of opposition to COVID restrictions and vaccine policies; it also seeks to increase coal mining.

Cartographically informed analyses of this election typically note that the imperial political boundaries that were imposed after the partition and annihilation of Poland in the late 1700s are still visible on the country’s electoral map. As The Economist’s article and accompanying map (see below) show, areas that were under Prussian (subsequently German) rule tended to vote for centrist, pro-EU parties, whereas those that were under Austrian and Russian rule were more inclined to support Euroskeptical, populist-nationalist parties. The same correlation was present in several earlier Polish elections. As The Economist explains:

[M]ost of the east belonged to tsarist Russia, where serfdom remained legal until 1861. By 1900 incomes in what is now western Poland were five times higher than in the east. This gap remains today: Poland’s four eastern provinces are all among the EU’s poorest 20 sub-national regions. Young people growing up in the east quickly move to larger cities, seeking education and private-sector jobs. Those who feel left behind have flocked to PiS, which offers both nationalist rhetoric and monetary hand-outs.

Such analysis is complicated, however, by the fact that most of the areas that had been under German rule had also been mostly populated by ethnic Germans. They were expelled after WWII, replaced mostly by ethnic Poles who had lived in the Russian-ruled east. The Economist explains this seeming paradox as follows:

The Soviet Union claimed a chunk of eastern Poland as the spoils of victory, while Germany was forced to relinquish its own eastern borderlands to Poland. The Polish government responded by resettling millions of people from the territory it lost to the areas it gained. Separated from their families’ fields and villages, these “repatriates” developed a more open and cosmopolitan identity, and grew less receptive to fist-thumping nationalism. Meanwhile, Catholicism remained strongest in Poland’s historic eastern heartland, which developed a fiery sense of pride and suspiciousness of change.


But while Poland’s former imperial divisions are an important factor in its current electoral geography, the situation is not a clear-cut as it might seem. The Economist’s featured map, for example, is based on administrative divisions at the powiat level; if one dives down to the more local gmina level, however, the correlation is no longer as obvious (see the maps below). Nor is The Economist’s economic generalizations about these former imperial divisions entirely accurate. As the Statistics Poland regional GDP map posted below shows, the Warmian–Masurian Voivodeship, formerly the southern half of Germany’s East Prussia, has Poland’s second-lowest level of per capita economic production. Another challenge to the imperial-legacy model of Polish electoral geography comes from comparisons of voting behavior and population density. A X-poster Thorongil notes – and maps (see below) – “a normal map shows the historical partition borders but the truth is that the opposition coalition’s most powerful vote centers are in Poland’s cities, big and small.”

As Thorongil’s map is difficult to interpret, I have tested his assertion by making a series of simpler maps. The first simply locates Poland largest cities on a detailed map of the 2023 election. As can be seen, urban gminy** gave a higher level of support to the centrist Civic Coalition than surrounding areas, but the correlation is by no means overwhelming. Comparing the electoral map to one of population density allows more precise assessment. To do so, I extracted spatial information from a detailed density map (below) and overlaid it on the 2023 electoral map; I also added dotted lines to roughly show the old imperial divisions. The first of these maps outlines Poland’s most sparsely inhabited areas, those with fewer than 50 persons per square kilometer. As can be seen, many of these areas in the former German zone supported United Right, seemingly upholding Thorongil’s density thesis over the imperial-legacy model. The next map outlines high- and medium-high density areas. As can be seen, high-density areas, marked with red and pink borders, had relatively high levels of support for the political center, but this linkage is stronger in the German zone than in the former Russian and Austrian zones. The correlation is not as close, however, in areas of medium-high density (200-100 persons per square kilometer). As can be seen, such areas surrounding the city of Poznan in the former German zone voted strongly for centrists, but most of those in the Katowice area, also in the former German zone, supported the United Right. A number of medium-density gminy in the Krakow (former Austrian) and Warsaw (former Russian) regions also supported United Right.

Both population density and imperial legacies were important factors in the 2023 Polish parliamentary election. But the situation is more complicated than it might appear, and other issues, many of which are of a more local nature, must also be considered. The next post will try to tease out some of them.

** This post considers only the election returns for Poland’s Sejm, its more powerful house of parliament, ignoring the senate vote.

** “Gminy” is the plural form of “gmina.”

Former Imperial Boundaries and Population Density in Poland’s 2023 Election Read More »

Neighborhood Stereotypes and Recent Voting Patterns in Auckland, New Zealand

Today’s post employs an unusual strategy for analyzing electoral geography, that of comparing local election results with neighborhood stereotypes. Here we look at the Auckland vote in New Zealand’s 2023 election, doing so in light of popular perceptions of different parts of the city as revealed by a detailed “judgmental map of Auckland” (published in 2017 in Newshub; see the previous post). To make the comparisons easier, I have overlaid maps of the 2023 election on sections of the stereotype map. My analysis of these combined maps is merely suggestive and is not informed by any firsthand knowledge of the city. It should thus be taken with a grain of salt.

We begin in the heart of the city, Auckland Central. The stereotypes of this Green-voting and strongly left-leaning electorate reflect its division into relatively rich and poor areas, an unexceptional feature for a central-city location: “expensive dining,” “hipsters,” “cruise ships,” “shows,” “porn,” “student ghetto,” “intensification,” and “done up.” The only term that I find confusing is “done up.” From a quick investigation of the term’s use in New Zealand, I infer that in this context it means “refurbished” or perhaps even “gentrified.” If this interpretation is correct, it is not a surprising designation for a Green-supporting area. As the map of the Green party-list vote in Auckland shows, support for the party is strongest near the urban core and declines in the peripheries (ignore the essentially unpopulated western expanse of New Lynn on the map).

The two Auckland electorates that selected candidates in the libertarian-leaning ACT Party, yellow-shaded Epsom and Tamaki, are affluent, inner-suburban communities. On the stereotypes map, Epson is prominently labeled “Double Grammar Zone,” a term that I originally thought might refer to a pretentious manner of speaking found among its well-off residents. Actually, the term is much more prosaic:

Three magical words significantly inflate the value and appeal of an exclusive group of Auckland residences. “Double Grammar Zone” is a most alluring catch-cry to many in this already searing hot market. Yes, these words offer the chance of access to the prestigious and successful Auckland Grammar School and Epsom Girls Grammar School. Experts have long pointed to the difference in price for properties located within the DGZ and noticed that gap widening as the city’s average house price continues to hit monthly record highs.

It is not surprising that residents of such an affluent neighborhood would support an anti-populist, economically conservative party. As the party-list vote map (below) shows, support for the ACT Party is highest in the city’s wealthy eastern fringe. In Epsom, the neighborhood labeled “professors” seems out-of-place; I can only assume that the voting patterns of this area are more like those of neighboring Auckland Central. At first glance, I found the “wankers” label mystifying, as I only understood this word as British term of general abuse that that originally denoted “masturbators.” The Urban Dictionary, however, claims that in New Zealand, Australia, and the UK, “wanker” primarily means “someone excessively and annoyingly pretentious and/or false, with a strong likelihood of working in the creative industries, especially ‘new media.’” I would not expect such “creative types,” however, to vote libertarian; perhaps this “wankerish” part of Epsom also has a different voting profile than the rest of the electorate. The label “Jon Ken” is even more mystifying; all that I could find when searching for that name was a nurse at an Auckland hospital.

The other ACT-supporting electorate in 2023, neighboring Tamaki, is labeled with several terms signifying establishment-oriented affluence: “yuppies,” “old-school suburbia,” “quite nice,” and “almost as nice.” The large-font label “Hannover Finance” refers to “a New Zealand non-bank finance company that focused on lending for high-risk property development that failed in 2010…” Its inclusion and prominence on the map perhaps reflect the common concern in Auckland about surging property prices. One label seems out-of-place for affluent Tamaki: “The Projects.” This term calls to my mind urban redevelopment initiatives in downtrodden neighborhoods. But the economic gradient between wealthy Tamaki and poor, Labour-voting Panumure-Otahutu (labeled “P-O” on the electorate map) to its south is steep, leading me to wonder whether the label has been placed a little too far to the north. But then again, the mapmakers have vastly greater knowledge of Auckland than I do.

The only electorate in the core region of Auckland that supported Labour in 2023 is Mount Albert. Its tags on the stereotype map suggest a relatively poor and ethnically diverse area that is changing as younger and more affluent people move in (“more hipsters,” “coffee,” “next to be gentrified,” and “halal.”) Such a district would be expected to heavily support both Labour and the Green Party, and that is exactly what one finds (see the map for the Green Party-list vote). I am confused, however, by the “Butcher’s” label in northern Mount Albert; perhaps it refers to upscale Omak Meats.

Two electorates in the southern part of the Auckland isthmus, Mount Rosekill and Maungskiekie, voted strongly for Labour in 2020 but switched to the National Party in 2023. Some of the labels placed here suggest a stable working- and lower-middle-class social environment: “alright suburbs,” “shabby suburbs,” and “panel beaters” (car-repair shops). “Mecca” and “noodles dumplings” probably indicate concentrations of immigrants from the Middle East and East Asia, respectively. (“McGehan Close,” to the contrary, denotes a street noted for its “hopelessness,” but it is located in Mount Albert, not Mount Rosekill, indicating either an error by the cartographer or one by me when I combined these maps.)

Labour’s main Auckland stronghold is in the southern part of the city (see the maps below). This is a decidedly poor and ethnically diverse area. The stereotype labels here are telling: “cleaners at your office,” “hardcase,” “Apia” (the capital of Samoa), and, in large font, “extra police resources.” The stark “extra police resources” tag, however, also extends into a much better-off electorate (Takanini), which switched from Labour to the National Party in 2023. To the north of Takanini are two relatively well-to-do electorates that have long supported the National Party and shunned Labour. One of them, Pakuranga, also has relatively high levels of support for ACT. Some of the stereotypes for this electorate, such as “bratty teens” and “wealthier bratty teens” are interesting, but I am especially intrigued by “paranoid South Africans.”

West Auckland (see the map below) includes another electorate that supported Labour in 2023, Kelston, although it did so by a relatively thin margin. Some of its stereotypes – such as “P-Labs” (meth labs) and “Tongans” – indicate the presence of rough neighborhoods and of a large Polynesian immigrant community. To its north is Te-Atatu; noted for its low- and medium-cost housing. The prominent label “Cheryl West” found here refers to a character in a popular television show who supposedly typified the “Westie” personality, defined by Wikipedia as someone “from the outer suburbs who [is] unintelligent, undereducated, unmotivated, unrefined, lacking in fashion sense, working-class or unemployed.” The same article, however, also notes that “Westie” has been gradually shifting from a “pejorative to a societal identifier,” based mostly on its prominence in television shows, song lyrics, and comedy routines. The movement of such a working-class redoubt as Te-Atatu from the Labour Party to the National Party in 2023 is therefore of some significance.


Another western electorate that switched from Labour to National in 2023 is New Lynn. Based on the stereotypes applied to it, such results are surprising. Such tags as “faint whiff of pot,” “hippies,” “potters,” and “artisany type people,” would suggest a decidedly left-leaning population. And that is its historical norm. As the non-updated Wikipedia article on the electorate notes, “It has always been held by members of the Labour Party.” But in 2023, the National Party triumphed in New Lynn both in the party-list vote and the electorate vote, albeit by relatively thin margins. The Green Party vote, however, was fairly large in New Lynn, as would be expected from the labels applied to it. Intriguingly, its new MP, Paulo Reyes Garcia, is an immigration lawyer originally from the Philippines.

The five northern electorates of Auckland (see the map below) all favored the National Party in 2023 in both the part-list vote and the electorate-based vote. In contrast, in 2020 all of them favored the Labour in the party-list vote, as did two in the electorate-based vote. The eastern part of this area is notably affluent, as is reflected by its stereotypes, and therefore would be expected to support the National Party. Two of these tags, “decile 10” and “more like decile 8-9,” need an explanation for non-New Zealanders; “decile” refers to a school-ranking system based on the socio-economic characteristics of their students, with “decile 10” denoting those in the top 10 percent. The term “Lorde” might also be mystifying for some people in other countries; it is the stage name of the well-known Kiwi musician Ella Marija Lani Yelich-O’Connor, who was raised in the Northshore electorate in the area under her name label.

The southwestern part of northern Auckland, the Northcote and Upper Harbour electorates, is a mid-income area noted for its Asian immigrants. Such features are indicated by three prominent labels on the stereotype map: “very average,” “Koreans,” and “Chinatown” (although Northcote also includes an area that is evidently populated by “artists too cool for cityside”). Upper Harbour, with its “depressing suburbs,” “car yards,” and “Koreans” saw a particularly sharp drop in support for Labour from 2020 to 2023.

This cursory analysis suggests that New Zealand’s National Party currently now enjoys a fairly broad level of support, extending well beyond its upper-middle-class base. It will be interesting to see whether it will be able to retain working-class and immigrant support in the coming years.

As a final note, in doing research for this post I was also surprised to learn that people from Auckland are often disparaged by other New Zealanders. As the Wikipedia article on the term “Jafa” notes:

Jafa is a slang term (sometimes pejorative  for a resident of Auckland, New Zealand. It is an acronym, standing for Just Another Fucking Aucklander. [I]t is considered to be representative of the boorishness of Aucklanders, or the envy of the rest of New Zealand, depending on the perspective. The term has wider currency than the earlier derogatory term “Rangitoto Yank.” A variant is Jaffa, Just Another Fuckwit From Auckland.  … Auckland is alleged to be full of rude, greedy and arrogant people, having a similar reputation as Mumbai and Kolkata in India, Milan and Rome in Italy, Paris in France, London in the United Kingdom, New York City in the United States, or Moscow and Saint Petersburg in Russia.  …Auckland is alleged to be a culturally alien place due to the much higher proportion of non-Maori and nonwhite populations than the rest of the country. Percentage-wise, Auckland has the seventh largest ethnic Chinese population among all urban areas outside Greater China. In the 2006 census, Asians comprised 18.9% of Auckland’s population but only 7.9% in Christchurch, and 14.4% of Auckland’s but merely 2.8% of Christchurch’s population are Pacific Islanders. Most new immigrants to Auckland are from East Asia and South Asia, while people immigrating to other parts of the country show higher percentage rates of UK and South African origins. Auckland is finding itself increasingly marginalised on sports traditionally identified with New Zealand culture, such as rugby and netball, because of high immigrant numbers from countries with little tradition of such sports.

Neighborhood Stereotypes and Recent Voting Patterns in Auckland, New Zealand Read More »

Auckland’s Electoral to the Right – and Comparisons with U.S. Cities

As noted in the previous post, the Auckland metropolitan area, like New Zealand as a whole, experienced a significant electoral shift to the right in the 2023 election. This swing is glaringly evident in the party-list vote (see the previous post for an explanation of this term). In 2020, the Labour list triumphed in 19 of Auckland’s 20 electorates; in 2023, its count was reduced to five, with the center-right Nationalist Party winning the other 15. Even in its strongholds, Labour’s vote percentage dropped substantially. The separate vote for electorate-based MPs in 2023, however, exhibited much more diversity, as is generally the case. Although Labour still took only five Auckland seats in that contest, the Green Party took another, that of Auckland Central. As the libertarian-leaning ACT Party won two Auckland seats in the electorate-based vote, the National Party’s overall take was 12, but that was twice as many as it took in 2020.

By far the largest city in New Zealand, Auckland is noted as well for its ethnic diversity. It reportedly has the “the fourth largest foreign-born population in the world, with 39% of its residents born overseas” (see the table below).  By the current standards of the United States, it is remarkable for such a large and ethnically diverse city to support a conservative political party. In the U.S., candidates from the left-leaning Democratic Party almost always come out ahead in both urban cores and inner suburbs, with support for conservative candidates mounting only as one moves into the outer-suburban and exurban belts. Although strongest in northeastern and Pacific-coast cities, this pattern is evident to some degree across the country. It can be seen, for example, in Houston, Texas (see below), which is one of the country’s most ethnically diverse large cities.This electoral disparity between New Zealand and the United States is to some extent a reflection of the different characteristics of the two country’s major political parties. It is also linked to the different political environments created by a mixed-proportional parliamentary system of government (New Zealand), which encourages minor parties, as opposed to a “winner take all” system (the United States), which encourages a two-party duopoly.

In the United States, the Republican Party has veered in a populist-nationalist direction since 2016, which has reduced its support in affluent suburbs while increasing it among non-metropolitan and working-class voters. In New Zealand, the populist-nationalist political space has long been occupied by the marginally successful New Zealand First Party. Although situated on the right on social and cultural issues, New Zealand First leans far enough to the left on economic matters for it to have joined a minority coalition government with the Labour Party in 2017. New Zealand’s National Party, with a platform based on “free enterprise, reduction of taxes, and limited state regulation,” is more similar to the American Republican Party before the Trump revolution of 2016 than it is to the Republican Party today. In some respects, the National Party has a more centrist orientation than the Republicans Party has had since the Eisenhour era of the 1950s. Because of such moderation, New Zealand’s ACT Party has staked its ground further to the socio-economic right, highlighting its firm commitment to “classical-liberal and small (or limited) government principles coupled with what the party considers as a high regard for individual freedom and personal responsibility.” All things considered, the relatively centrist orientation of the National Party allowed it to take advantage of the current discontent with the policies of the out-going Labour government, and thus score a decisive victory.

As the various electorates of Auckland saw markedly different electoral results in 2023, it would be useful to see how well the returns correlate with demographic and socio-economic indicators. Let us begin with population density, which was a factor in the national vote (see the previous GeoCurrrents post). In Auckland, however, the role of population density is relatively modest – much less than what one would expect in the United States. As can be seen in the two paired maps posted below, the higher-density areas of the Auckland region generally showed a higher level of support for the Labour and Green parties than did lower-density areas, but the linkage is relatively small. A few relatively high-density areas, moreover, strongly supported the National Party.

Much closer correlations are found for economic class, which is evident in the paired maps posted below. The household income level map on the left is admirably detailed, but it has neither a key nor a textual explanation in its accompanying article; one can, however, deduce that the blue dots indicate high-income levels and the red dots low-income levels. By comparing the two maps, we can see that affluent areas tended to supported candidates in the National Party, whereas the richest ones were more supportive of the libertarian-leaning ACT Party. In contrast, the poorest areas, located in southeastern Auckland, overwhelmingly supported the Labour Party. Such a pattern is reminiscent of that found in the United States decades ago. In an American city like San Francisco, however, the correlation between economic class and voting behavior in national elections has almost entirely collapsed (see the maps below).

In Auckland, some local deviation is found in such class-based voting behavior. The most left-leaning electorate, “green” Auckland Central, includes some notably wealthy areas, although it also has some relatively poor ones. This is not surprising, however, as people who vote for green parties across the world tend to be highly educated and relatively affluent. More surprising is the switch from the Labour Party to the National Party in several relatively poor parts of western Auckland, as well as in some of the more economically mixed areas of north Auckland.

Such patterns deserve further scrutiny, which is difficult to accomplish for someone with limited time, no assistance, no personal knowledge of the city. Fortunately, however, Newshub has published a fascinating and detailed map of neighborhood stereotypes in Auckland that might prove useful in this regard. The next GeoCurrents post will therefore see whether such a “judgmental” view of the city can shed any light on its recent electoral shift.

Auckland’s Electoral to the Right – and Comparisons with U.S. Cities Read More »

New Zealand’s Striking Electoral Shift to the Right

The conservative National Party of New Zealand scored a major victory in the country’s October 2023 general election, with the governing Labour Party suffering a historic defeat. As described by The Guardian, “New Zealand voters have delivered a forceful rejection of the Labour government as a surge in support for the National party delivered what analysts described as a ‘bloodbath, for the government and a new right-leaning era for politics in the country.” But just three years earlier, it was Labour in the victory circle, winning the 2020 election so overwhelmingly that it was able to govern without a coalition partner. But in the intervening period, the country’s mood soured over concerns about high taxes, increasing crime, the rising cost of living (especially of housing), and the government’s highly restrictive COVID policies.

Before delving into geographical analysis of New Zealand’s recent elections, it is necessary to explain the complexities of the county’s “mixed-member proportional” parliamentary system. New Zealand is divided into 65 general “electorates” (geographical voting constituencies) and then redivided into seven special electorates for Māori voters. Each electorate selects one person to serve as its MP (Member of Parliament) in the unicameral parliament, officially known as the New Zealand House of Representatives. But Kiwi voters not only choose an individual to represent their electorate, but also vote a second time for a political party, each of which maintains a list of potential MPs. Parties whose total vote in that contest exceeds a certain threshold (usually five percent) send an additional 48* MPs into the House of Representatives, their numbers proportional to their share of the vote. Minor parties can thus gain parliamentary representation either by having enough voters concentrated in one or more electorate to defeat candidates from the other parties, or by having enough support nationwide to crack the five-percent threshold.

Labour’s overwhelming triumph in the 2020 election is strikingly evident on the map of the “party list vote,” which is on the left side of the paired Wikipedia maps posted below. Astoundingly, the Labour-list came in first place in all but one general electorate. Its rival center-right National Party took only a single district, located in a suburban area of Auckland. The direct electorate results were much more balanced, with individual candidates in the National Party taking seats in both non-metropolitan areas and in the more affluent parts of Auckland (see the map on the right). Three other parties – the Green Party, the Maori Party, and the “classical liberal” ACT Party – also sent MPs to parliament in 2020, based both on their national party-list vote and on their victories in individual electorates. All in all, 2020 was a banner year for New Zealand’s political left, with Labour, the Green Party, and the Maori party (Te Pāti Māori) together holding 78 parliamentary seats, as opposed to 42 held by the center-right National and ACT parties.

On October 14, 2023, however, New Zealand experienced a stunning electoral reversal. As the party-list vote maps for the two elections show, New Zealand went from almost entirely red (Labour) to almost entirely blue (National Party). Even on the more diverse Wikipedia map of the direct electorate results, there is little red to be seen in the country as a whole. But such mapping is misleading; as the inset maps show, the Labour and Green parties won quite a few urban seats, particularly in the country’s second and third largest cities, Christchurch and Wellington. But overall, the 2023 election was a clear triumph for conservatives. It was also a rout for Labour, which went from 62 to 34 seats in the House of Representatives. But the other left-leaning parties, the Greens and the Māori Party, gained seats. So too did the classically liberal ACT Party. The socially conservative nationalist-populist New Zealand First Party also did relatively well, returning to the House of Representatives after an absence of several years.

As conventional electoral maps give undue prominence to sparsely inhabited areas, and therefore tend to visually exaggerate the vote-share of conservative parties, electoral cartographers have devised more representative maps. The usual strategy is to expand more densely populated areas in proportion to their populations. For New Zealand’s 2023 election, The Spinoff devised such a map, converting the country into hexagons of roughly equal population. It also grouped the parties into two categories, one left-leaning and the other right-leaning. As can be seen in the resulting map, in the 2023 election New Zealand was still a mostly blue (conservative-voting) country, although not to the extent seen in conventional maps. This Spinoff map also clearly shows the Māori population, with its special electorates, as strongly supporting the political left.

The Spinoff has drafted another map that divides New Zealand’s electoral hexagons into three categories, one composed of large cities (Auckland, Wellington, and Christchurch), one of medium-sized cities, and one of rural areas. Such mapping helps us see the role of population density in voting behavior. To clarify this situation, I have “whited-out” non-rural areas on one iteration of this map, everything but large cities on another, and everything but medium-sized cities on a third. As can be seen, rural electorates supported the conservative National Party, although some by relatively thin margins. Medium-sized cities delivered more mixed results, with some strongly favoring the National Party and others supporting Labour. Dunedin, in southeastern South Island, in particular leans left. Such affiliation is strongest in North Dunedin; as “Just Dave” comments in a Quora query about New Zealand’s most left-wing cities:

The cities in which the most left-wing party that actually gets elected to Parliament (the Greens) receives the largest proportion of the popular vote in the are central Wellington, central Auckland and north Dunedin. All three areas have a comparatively young, wealthy and educated population. North Dunedin is primarily home to university students and university staff, for example.

Surprisingly, New Zealand’s large cities also appear as politically mixed on The Spinoff’s 2023 electoral map. To be sure, Wellington – the capital – is mostly red (Labour) and green (Green), but it is a different story in Christchurch and especially Auckland. Auckland, by far the largest metropolitan area in the country, deserves a more detailed analysis – which it will receive in the next GeoCurrents post.

*This number can be slightly higher due to extenuating circumstances.

New Zealand’s Striking Electoral Shift to the Right Read More »

Small But Densely Populated American Cities & the Transformation of Cudahy, CA

The list of the most densely populated incorporated cities in the United States has some interesting features. The top four entries are all small cities (less than 1.5 mi sq; fewer than 70,000 inhabitants) located just to the west of Manhattan in Hudson County, New Jersey. Three of the top 11 – Kaser, New Square, and Kiryas Joel – are relatively new towns in the New York metropolitan area that are entirely or primarily inhabited by Hasidic Jews. All three have high fertility rates and low levels of per capita income. According to Wikipedia, “Kiryas Joel has the highest poverty rate in the nation” while New Square is “the poorest town (measured by median income) in New York, and the eighth poorest in the United States.”

One surprising revelation in the city-density list is the large number of thickly populated cities that were originally established as low-density suburbs of Los Angeles. Of the 140 U.S. cities with more than 10,000 people per square mile, 28 are in the Los Angeles region. Although still conventionally imagined as a low-density, suburban environment, the L.A. region has been densifying for decades. The sprawling city of Los Angeles itself, covering some 469 mi sq, is now moderately dense by U.S. standards. As the density map of southern Los Angeles County posted below shows, central L.A. is now heavily inhabited, with many census tracts reporting more than 30,000 people per mi sq. Quite a few outlying tracts also post high figures. Many of these areas do not appear at first glance to be densely populated, as they are dominated by low-rise buildings and include many detached, single-family houses. But the number of persons living in each dwelling unit can be high, particularly in areas with large numbers of recent migrants.

Several of small, densely populated cities in the Los Angeles metropolitan area in the northwestern quadrant of a cluster of municipalities known as the “Gateway Cities.” I have enclosed the northern portion of this “Gateway” area on maps posted above and below, excluding the relatively large city of Long Beach. The crowded little cities in this region are relatively poor and have large immigrant populations. In 2019, Business Insider placed Huntington Park in the lowest position in California on its “misery index” and in the tenth lowest nationally. The Wikipedia article on Maywood estimates that one-third of [its] residents live in the U.S. without documentation.” Maywood is also notable for being “the first municipality in California to outsource all of its city services, dismantling its police department, laying off all city employees except for the city manager, city attorney and elected officials, and contracting with outside agencies for the provision of all municipal services.”

The evolution of tiny but densely packed Cudahy, with almost 23,000 residents living in 1.18 mi sq, is particularly interesting. Cudahy was originally designed as a semi-rural garden city. Its founder and namesake, the wealthy meat-packing entrepreneur Michael Cudahy, purchased a large ranch in 1908, which he subdivided and sold off in one-acre lots. As explained in the Wikipedia article on the city:

These “Cudahy lots” were notable for their size—in most cases, 50 to 100 feet (15 to 30 m) in width and 600 to 800 feet (183 to 244 m) in depth, at least equivalent to a city block in most American towns. Such parcels, often referred to as “railroad lots,” were intended to allow the new town’s residents to keep a large vegetable garden, a grove of fruit trees (usually citrus), and a chicken coop or horse stable.

Although gardens, orchards, and farm animals are long gone, the old “Cudahy lots” may still be visible in satellite images (see the image below; I was not, however, able to find a map of the original city lots). At any rate, Cudahy gradually morphed into a crowded industrial town, giving it a legacy of environmental contamination. As noted by the Wikipedia article cited above:

On January 14, 2020, delta Airlines flight 89 dumped jet fuel  Cudahy, while making an emergency landing at Los Angeles International airport. Park Avenue Elementary School suffered the brunt of this dumping. This incident sparked outrage because of the city’s previous history of environmental damage, including the construction of the same school on top of an old dump site that contained contaminated soil with toxic sludge, and pollution from the Exide battery plant.

As a final note, it is intriguing that the two main clusters of small, high-density cities in the United States are located immediately adjacent to the country’s two largest cities, New York and Los Angeles. Populous though they are, these two cities have markedly different built environments and settlement histories. New York is well known for its high population density, but Los Angeles is more commonly regarded as a low-density city anchoring an even lower-density metropolitan area. That vision is longer justifiable.

Small But Densely Populated American Cities & the Transformation of Cudahy, CA Read More »

Capturing the Size and Density of New York City and Environs on a Map of Major U.S. Cities

As mentioned in the previous post, depictions of the population density of major U.S. cities tend to under-emphasize the significance of New York City. New York is clearly the most densely inhabited major city in the United States, with 29,303 people per mi sq (in 2020), a figure that far overshadows that of second-place San Francisco (18,631). San Francisco, moreover, makes a poor comparison, as its total population is more than an order of magnitude less than that of New York (808,437 vs. 8,335,897 in 2022).

The population concentration found in the core areas of New York City is also masked by the relatively low density of some of its outlying areas, particularly of Staten Island. With a population of 8,618 per mi sq (in 2020), Staten Island is comparable in this regard to Los Angeles (8,304.22 per mi sq). In contrast, Brooklyn – which would be the country’s second most populous city if the boroughs of New York had never amalgamated – had a population density of 39,438 per mi sq in 2020, a far higher figure than that of San Francisco. But it is Manhattan that really stands out. Its 1,694,251 residents (2020) are crowded into a mere 22.83 square miles, giving it a density of 74,781 people per sq mi. A century earlier, Manhattan had been even more densely populated. When its population peaked at 2,331,542 in 1910, its density exceeded 100,000 people per mi sq, a figure that makes San Francisco seem sparsely settled in comparison.

In short, when it comes to both urban population size and density in the United States, New York City is in a league of its own, with no real competition. To illustrate this situation, I have redrafted two of the maps that were used to illustrate the previous GeoCurrents post. In the new versions (below), New York is broken down into its five constituent boroughs. A new density scheme was required as well, as four of New York’s five boroughs monopolize the top three categories in the new 2022 map. As the redrafted 1950 map shows, Queens and especially Staten Island were much less densely inhabited than the other boroughs at the time. This map highlights the significance of Brooklyn, the Bronx, and especially Manhattan as the country’s most densely populated urban places in the mid-twentieth century.

But even this redrafted map does not adequately capture the elevated population densities found in the greater New York City region. As the table of the most densely populated incorporated cities in the United States (posted below) reveals, New York City itself ranks in only the sixth position. The four cities with the highest density are all in Hudson County, New Jersey, immediately to the west of Manhattan. The largest city in Hudson County – Jersey City – is not on this list. But if cities that cover very small areas (below five square miles) are excluded, Jersey City ranks in the second position. Yonkers, which is immediately north of the Bronx, also makes this list of the most densely populated sizable U.S. cities. To reflect this concentration of dense urbanism in the New York metro area, I have edited the map once more, this time including Hudson County and Yonkers.


One more GeoCurrents post will examine population density in American cities. After that, this blog will turn to the recent elections in New Zealand and Poland before returning to the historical development of the urban system of the United States.

Capturing the Size and Density of New York City and Environs on a Map of Major U.S. Cities Read More »

U.S. City Size, Density, & Population Change, 1950 to 2022 – and the Dream of the “15-Minute City”

Many environmentalists now advocate the development of “15-minute cities,” urban areas dense enough to allow residents “to access most of the places [they] need to go within a 15-minute walk or bike.” This vision has much to recommend it. Many people find neighborhoods of this sort deeply attractive, both as places to live and visit. I count myself among them. My ideal living arrangement would be to divide my time between an apartment in such a city and a house in a remote rural area. Instead, like most Americans, I live in a medium-density suburban environment – which sometimes seems to offer the worst of both worlds.

But although I understand the appeal of 15-minute cities, I also recognize that creating them would be extraordinarily difficult if not impossible in the United States. Evidence from both polling and actual residential choice indicates that most Americans dislike dense cities and prefer suburban living. Ironically, moreover, environmentalists themselves are one of the main obstacles to the urban intensification that such a vision requires. Construction projects of all sorts, after all, often face environmental lawsuits, which can bring them to a quick halt.

An equally severe problem is the fact that the few cities in the United States that approach the required degree of walkability have been deintensifying, shedding residents over the past several years. From 2020 to 2022, New York City lost 3.5 percent of its population, Philadelphia 2.3, Chicago 3.0, and San Francisco a shocking 7.5. This decline was at first mostly a matter of people fleeing crowded conditions during the COVID pandemic, but it is now being driven primarily by safety and property-security concerns. For the same reasons, many of the mass-transit systems that are required for urban intensification are losing ridership and find themselves financially troubled. As a result, it is difficult to avoid the conclusion that, at least in the United States, the 15-minute city is little more than a fond dream.

Some of the maps that I have been making for my prospective historical atlas of urban development in the United States might prove useful in examining the urban growth and density issues surrounding the 15-minute city idea. These maps, to be sure, are unusual, as they depict no geographical features beyond city size and density. The spatial patterns that they show are also wildly distorted. As a result, they might more properly be regarded as graphic visualizations. But I still view them as maps, as all GeoCurrents posts focus on map explication.

The first map shows the size, density, and rough relative locations of the twenty most populous cities, as formally defined, in the United States in 2022. The numbers in the bottom corners of each urban polygon indicate the population growth rate, in percentage terms, of that city from 2010 to 2020 (left) and from 2020 to 2022 (righ). As can be seen, most large American cities lost population in the latter period. More important, such losses were concentrated in more densely inhabited cities. Several of the more sparsely settled cities, in contrast, gained population during this period. But as can also be seen, all these cities added residents from 2010 to 2020, some of them to a significant degree. This was true even in the country’s most densely inhabited urban areas. New York grew in this period by 7.7 percent and San Francisco by 8.5 percent. But with the exceptions of Seattle and Denver, all cities expanding by more than ten percent from 2010 to 2020 are characterized by low population density.

The overall impression conveyed by this map is one of low population density in America’s largest cities. Some of them have annexed such extensive suburban and rural hinterlands that they do not really count as cities in the informal sense. Jacksonville, Florida, for example, consolidated with Duval County in 1968, and as a result, its 971,319 residents live in a “city” that sprawls over 874.46 sq mi. This gives Jacksonville a population density of 1,270.73/sq mi, a figure lower than that of the typical American inner suburb. The contrast between Jacksonville and San Francisco is instructive. Although the city of San Francisco is also consolidated with its county, its population density is of an entirely different magnitude. In 2022, San Francisco’s 808,437 residents inhabited an area of 46.9 sq mi, giving it a density of 17,237.5/sq mi. But if San Francisco is thickly populated by U.S. standards, it is not by that of New York City. In 2020, Manhattan had 1,694,251 residents living in an area of 22.83 sq mi, giving it a density of 74,780.7/sq mi.

As the next map shows, in 1950 the 20 largest cities in the United States were considerably denser that those of 2022. 1950 was arguably the heyday of American urbanism. Driven in part by the war-economy of the first half of the decade, all large U.S. cities grew during the preceding census interval, some by considerable margins. Extremely rapid growth occurred both in sparsely inhabited cities (see Houston on the map below) and in densely settled ones such as San Francisco and Washington, DC.

Seven cities are found on the lists of the 20 largest U.S. cities in both 1950 and 2022. As can be seen on the map posted below, the country’s two densest major cities, New York and San Francisco, experienced relatively little change in either population size or density in the intervening 72 years. Two relatively densely settled cities, Chicago and Philadelphia, saw significant populations losses in the same period, reducing their densities. In contrast, two West Coast cities, Seattle and Los Angeles, experienced major increases in both population and density. Houston, in contrast, saw a huge population increase but did not more into a higher population-density category, as it also expanded in area.

The next map, indicating population size but not density, shows which cities dropped out of the top-20 list between 1950 and 2022 and which ones were added to it. The geographical pattern seen here is stark but not surprising. Except for New Orleans, all the “drop-out” cities are in the northeastern quadrant of the country. In contrast, with the exceptions of Indianapolis and Columbus, all the additions are in the southern half of the country. Interestingly, Columbus has many attributes of a sunbelt city, although it experiences very little sunshine from November through March. The concentration of emergent, low-density, large cities in Texas is also noteworthy.

The final map addresses a question that probably crossed the minds of some readers: where are such major cities as Atlanta or Miami? With just under half a million residents, Atlanta is not a particularly large city, although its metropolitan area certainly is. The same patterns holds for Miami. The map below thus shows the locations (but not the populations) of cities that anchor metropolitan areas in the top 30 by population in 2022, but did not themselves place in the top-20 city lists of either 1950 or 2022. It is not coincidental that three of the eight are in booming Florida.

The first two maps in this post are somewhat misleading, as they do not adequately convey the population density of New York. To do so properly, the city must be broken down into its five constituent boroughs. This will be done for the next GeoCurrents post.

U.S. City Size, Density, & Population Change, 1950 to 2022 – and the Dream of the “15-Minute City” Read More »

Mapping the Development of the Urban Framework of the United States, 1790-1830

I am currently working on an online historical atlas of the development of the urban framework of the United States. The maps and commentaries that will constitute this atlas will be posted gradually over the next few weeks or months, interspersed with regular GeoCurrents posts. The first of these installments, showing the situation in 1840 and outlining the “Philadelphia problem,” appeared on October 13, 2023. Today’s post examines the development of the network of cities in the United States from 1790 to 1830. The population figures in today’s post, like that of October 13, are derived from a Wikipedia article called “List of Most Populous Cities in the United States by Decade.” In subsequent posts, covering the period after 1840, a more comprehensive data source will be used.

The United States had few cities of any size in 1790. New York City tops the conventional list, with 33,131 inhabitants, and Philadelphia comes in second, with 28,522. But Philadelphia at the time was limited to what is now called Center City. If one includes what were then the separate cities of Southwark and Northern Liberties District, which were annexed in 1854, Philadelphia ranks first, with a population of 44,096, and is mapped accordingly.  As can be seen on the map posted below, the country’s main cities – or towns, in one prefers – of the time were all ports, located on the coast or along estuaries. Except for Charleston, South Carolina, all of them were in the greater northeast. The prominence of New England on this map, with more than half of the cities depicted, will not persist into the 1800s as the urban center of gravity shifts south into the Mid-Atlantic states.

The largest cities on the 1790 list significantly expanded from 1790 to 1800, with New York growing from 33,131 to 60,514, Baltimore from 13,503 to 26,514, and Boston from 18,320 to 24,937. Philadelphia, in the larger sense, still vies with New York for top position. Norfolk, Virginia appears on this map, but the year 1800 marks its only inclusion in the top-ten list.

The rapid expansion of the country’s largest cities is a persistent feature of these maps. By 1810, the population of New York City approached 100,000. By this time, New York was clearly the country’s largest city, a position that it will retain and amplify in the following decades. The 1810 map includes the first truly inland city, Albany, New York. Located on the Hudson River, Albany’s appearance reflects the growing importance of trade with the interior. More important is the inclusion of New Orleans on the southern Mississippi, which became part of the United States with the Louisiana Purchase of 1803.

In 1820, Albany drops of the map, replaced by Washington DC, which had 13,247 inhabitants in that year. But as the nation’s capital experienced relatively slow growth after this period, it falls off the top-ten list in 1830 and does not reappear until 1950. In the early nineteenth century, Washington was derisively called “the city of magnificent distances” due to its small number of residents living in an urban framework designed for a larger population. In 1842, Charles Dickens claimed that “Its streets begin in nothing and lead nowhere.” The fact that capital of the United States was such a small city reflects the limited extent of the federal government before the Civil War. As its constituent states were arguably more important than the country itself, the common locution at the time was “The United States are…,” rather than “the United States is… .”

The major changes on the map of 1830 reflect the opening of the Erie Canal (the dotted blue line on the map) in 1825. The Erie Canal facilitated the emergence of an extensive water-based transportation network, linking the Hudson River to the Great Lakes, and, by extension, to the Ohio and Mississippi rivers. Not surprisingly, Albany reappears on the 1930 map. More important, Cincinnati emerges as the first significant Midwestern city. Cincinnati will remain in the top-ten list until 1910. Today, with a population of 309,51, it ranks in the 64th position, surpassed by a few suburbs of little historical significance. In the early and mid-1800s, however, Cincinnati was a major and rapidly growing city, due in part to its role in butchering and processing hogs for the national market. This industry was so important that the city was deemed “Porkopolis.” As is explained in a 2016 Cincinnati Magazine article:

“Porkopolis” is one of the names by which Cincinnati is known, and its origin is explained in the following manner: About 1825 George W. Jones, president of the United States branch-bank, and known as “Bank Jones,” was very enthusiastic about the fact that 25,000 to 30,000 hogs were being killed in this city every year; and in his letters to the bank’s Liverpool correspondent he never failed to mention the fact, and express his hope of Cincinnati’s future greatness as a provision-market. The correspondent, after receiving a number of these letters, had a unique pair of model hogs made of papier mache, and sent them to George W. Jones as the worthy representative of ‘Porkopolis.’”

… Frances “Fanny” Trollope is infamous for publishing a scathing indictment of Cincinnati in her 1832 book “Domestic Manners of the Americans”. A great deal of her bile is directed at our pigs:

“If I determined upon a walk up Main-street, the chances were five hundred to one against my reaching the shady side without brushing by a snout fresh dipping from the kennel; when we had screwed our courage to the enterprise of mounting a certain noble-looking sugar-loaf hill, that promised pure air and a fine view, we found the brook we had to cross, at its foot, red with the stream from a pig slaughterhouse while our noses, instead of meeting ‘the thyme that loves the green hill’s breast,’ were greeted by odours that I will not describe, and which I heartily hope my readers cannot imagine.”

It is not coincidental that the Procter & Gamble Company is headquartered in Cincinnati. As explained in Encyclopedia Britannica:

The company was formed in 1837 when William Procter, a British candlemaker, and James Gamble, an Irish soapmaker, merged their businesses in Cincinnati. The chief ingredient for both products was animal fat, which was readily available in the hog-butchering centre of Cincinnati. The company supplied soap and candles to the Union Army during the American Civil War and sold even more of these products to the public when the war was over.

Although candles are now usually made of wax, historically they were mostly made from animal fat. In earlier times, only prosperous people could afford wax candles.

Mapping the Development of the Urban Framework of the United States, 1790-1830 Read More »

Mapping the Population of U.S. Cities in 1840 – and the Philadelphia Problem

I am currently working on a large set of GeoCurrents maps that will depict the current and historical demographic patterns of U.S. cities and metropolitan areas. Several problems, however, have arisen in data selection and visualization. Most troublesome is the gradual amalgamation of separate municipalities into single cities.

Consider, for example, a map showing the locations and populations of the six largest U.S. cities in 1840 (below). It might be surprising that Philadelphia, considering its historical importance, appears as only the fourth largest, surpassed in population by Baltimore and New Orleans. But this depiction is misleading.  As it turns out, 5 of the 37 American cities with more than 10,000 inhabitants in 1840 are now mere neighborhoods of Philadelphia. New York City, as it is currently conceptualized and legally defined, was also larger than it appears on the map.  In 1840, it did not include Brooklyn, which was then the country’s seventh largest city. Boston was larger as well, as it did not then include Charlestown (see the table below).

To address this problem, I have revised the map by amalgamating what were then separate municipalities with the nearby cities that later annexed them. I did so, however, only for cities with more than 10,000 inhabitants. If smaller cities were subjected to the same treatment, the map might have to be revised again. But regardless of such difficulties, it can be clearly seen that Philadelphia was the country’s second city in 1840, with a population more than twice that of Baltimore.

The final map includes all cities (that currently exist as cities) that had more than 10,000 inhabitants in 1840. As can be seen, almost all were linked to transportation networks, serving as ports on seacoasts, estuaries, or rivers. Several are located on the Erie Canal (shown as a dotted blue line), again illustrating the importance of waterways in the pre-railroad era, which quickly coming to an end. Lowell, in northeastern Massachusetts (mapped in a light shade of red), is an interesting exception, as it emerged as a planned industrial city focused on textiles. Located on the rapids of the Merrimack River, which provided power, Lowell is often regarded as the “cradle of the American industrial revolution.”

Mapping the Population of U.S. Cities in 1840 – and the Philadelphia Problem Read More »

Insurgency in Paraguay – and Genocidal Agitation Against Brazilians in the Country

Wikipedia’s “list of on-going armed conflicts” (see the previous post) had some surprises for me, as it includes a few insurgencies that I had thought were over. One example is that of the Paraguayan People’s Army, or EEP Rebellion (from the Spanish label, Ejército del Pueblo Paraguayo). Wikipedia gives a 2023 death toll of seven for this conflict, and a cumulative count of 145+ since its beginning in 2005. These figures do not seem to be reliable, however, as the listed source for the 2023 figure is from 2022. I was not able to find any information on deaths this year in an admittedly cursory internet search. The Wikipedia article on the EEP, however, emphasizes its continuing activity, claiming that it can field up to 1,000 militants. As the article notes:

[T]he EPP has millions of dollars collected in kidnappings, extortion, expropriations and even contributions from neighbors and supporters. To this day, they continue to gain followers in the area, given the void left by the Paraguayan State.

The EEP is in many respects a typical Latin American Marxist-Leninist insurgency. It aims its attacks on wealthy landowners and security official, both private and public. Its operations have been focused in the central-eastern part of the country not far from the boundary with Brazil (see the map below), a restive region that has seen the development of large, mechanized farms over the past few decades. A few years ago, the EEP gained some global notoriety for kidnapping Mennonite farmers, one of whom was killed when his family was unable to come up with the $500,000 demanded for his release.

Conflict over land use and ownership in eastern Paraguay is an issue of the political far-right as well as the far-left. In Paraguay’s April 2023 general election, the populist and self-described nationalist-anarchist candidate of the National Crusade Party, Paraguayo “Payo” Cubas, surprised many observers by coming in third place, taking almost a quarter of the votes cast. In 2019, then-senator Cubas was impeached after he called for the genocide of Brazilians living in his country. As reported by Folha de São Paulo:

Brazilian bandits, bandits! Invaders! Now deforesting the country,” he shouts. “At least 100,000 Brazilians must be killed here,” he continued, mentioning that 2 million Brazilians are living in the country. The Brazilian government estimates that there are 350 thousand.

Following his failed bid for the presidency, Cubas was arrested for “disturbing the peace” after he refused to accept the election results and led anti-governmental protests. This was not the first time that he found himself in legal trouble. In 2016, Cubas was arrested “after hitting a judge with a belt and defecating in the office of the judge’s secretary.”

The large Brazilian presence in eastern Paraguay dates to the 1960s. These so-called “Brasiguayos” (“Brasiguaios” in Portuguese), many of whom were born in Paraguay, are now thought to number around half a million, a little less than 10 percent of the country’s population. They form the dominant group in several border towns, which are now mostly Portuguese speaking. This fact is almost never noted on language maps of Paraguay, although I did find one somewhat dated example (posted below). This map, not surprisingly, comes from the extensive archives of Reddit’s “Map Porn” community.

The initial Brazilian immigrants in Paraguay were mostly landless peasants who cleared the land for agriculture. They were later followed by well-off farmers who developed mechanized, commercial agriculture, usually focusing on soybeans. As commercial farmers moved in, many of the earlier migrants were forced back to Brazil, where they often found themselves unwelcome. Settling mostly in the new agricultural areas of Matto Grosso do Sul, their plight gained the attention of Amnesty International, which claimed in a 1992 report that were the victims of “illegal detentions, allegations of excessive use of force by the police, intimidation and a possible extra judicial execution.” The irony inherent in the situation has been noted. As one author put it, “Brazilians living in Paraguay wound up being expelled by their own countrymen.”

Anti-Brazilian agitation in Paraguay over the past few decades has generally focused on landownership issues. It seems to have reached a peak between 2008 and 2012, when Paraguay was under a leftwing government, an unusual condition in that country. As noted in a 2012 article in Gazeta do Povo:

The epicenter of the most recent agrarian conflict in Paraguay is located 75 kilometers from Foz do Iguaçu, in the department of Alto Paraná. A group of 6,000 landless Paraguayans, called “carperos”, have been camped for almost a year in the municipality of Ñacunday, on the border between two rural properties owned by producers of Brazilian descent. They threaten to take by force an area of 167,000 hectares spread across the departments of Alto Paraná, Canindeyú and Itapúa on the border with Brazil and Argentina. Armed and willing to radicalize the movement, they claim that the lands occupied by Brazilians belong to the Paraguayan government and should serve the agrarian reform project undertaken by President Fernando Lugo.

Cultural and even racial issue are also at play. As reported in a 2001 New York Times article:

They complain that the only television available locally is Brazilian and that their children grow up rooting for Brazil’s national soccer team instead of their own and speaking Portuguese as their second language instead of the Indian language Guaraní [Note: Paraguay is almost completely bilingual in Spanish and Guaraní].

Radio broadcasts in Guaraní urging landless peasants to rise against the Brazilians continue to be heard here. About 80 percent of San Alberto’s 23,000 residents are of Brazilian descent, and by voting as a bloc they have succeeded in electing one of their number, Romildo Maia de Souza, as mayor. …

One source of friction, all sides agree, is racial. Many of the Brazilians are blue-eyed, fair-skinned descendants of the German, Italian and Polish immigrants who flocked to Brazil’s three southernmost states a century ago. Many of the native-born Paraguayans most resentful of the Brazilian presence are of [indigenous] Indian stock.

Finally, geopolitical implications further complicate the situation. A 2019 scholarly paper by Andrew Nickson warns that Paraguay might be a Brazilian “protectorate in the making,” which seem a bit exaggerated. A big up-coming issue in this regard is the renegotiation of the Itaipú Treaty, which covers the shared Itaipú dam, the third largest hydroelectric facility in the world.

Insurgency in Paraguay – and Genocidal Agitation Against Brazilians in the Country Read More »

Mapping Recent War Fatalities and the Persistence of Current Armed Conflicts

As noted in the previous GeoCurrent post, the civil war in Burma/Myanmar is one of the bloodiest conflicts in the world today. According to a comprehensive Wikipedia table, its death toll thus far in 2023 is 10,790, the fourth highest in the world. It follows only the Ukraine-Russia war (83,637-100,000+), the war in Sudan (11,501), and the multifaceted insurgency in the Maghreb/Sahel (10,868). Given the rapidly mounting number of fatalities in the current war between Israel and Hamas, however, the rankings for 2023 will probably have to be revised. In any event, in 2022 Burma had the third highest death count if one uses the upper range of estimates found in the table (20,206, as opposed to 109,600+ in Ethiopia and 100,000+ in Ukraine).

Burma’s civil war is also extraordinarily long-lasting, dating from 1948. The only on-going wars listed by the Wikipedia as having started earlier are the Kurdish insurgency in Iran (1918), the “Jamaican political conflict” (1943), and the insurgency in Kashmir (1947). The article also lists the Israeli-Palestinian conflict and the Baloch insurgency (in Pakistan and Iran) as having begun in the same year as Burma’s civil war, 1948. As of October 6, 2023 – when this post was initially written – none of these other armed conflicts had been nearly as deadly over the previous 10 months as that of Burma. On October 6, the Wikipedia table provided the following 2023 death tolls for these persistent conflicts: Kurdish insurgency in Iran, 147; Jamaican political conflict, 295; insurgency in Kashmir, 433; Israeli-Palestinian conflict, 287; and Baloch insurgency, 500. As of today, however, it gives the Israeli-Palestinian conflict a fatality count of 1,827. All told, if one combines recent death tolls and conflict duration, Burma’s civil war and the Israeli-Palestinian conflict seem to be the most serious conflicts in the world today.

The Wikipedia article under consideration includes a serviceable map of the “number of combat-related deaths in current or past year” (posted below). It might seem odd to place Mexico in the highest category (more than 10,000 fatalities), but the source includes “drug wars,” an intriguing but questionable move. As the map shows, wars today are concentrated in northern and central Africa, the Middle East, southern Asia, northern South America, Mexico, and Ukraine & Russia. In contrast, East Asia, Central Asia, Europe, southern Africa, southern South America, northern North America, and Oceania (Australia and the Pacific) are nearly free of armed conflicts.

This map, however, as well as the table that was used to generate it, must be regarded as highly approximate. It is extremely difficult, if not impossible, to consistently and accurately tabulate deaths in armed conflicts. Although the Wikipedia article does an admirable job, it misses some deadly conflicts. It does not list Papua New Guinea, for example, as having experienced any combat-related fatalities over the past two years. In actuality, so-called tribal wars in New Guinea’s highlands are ubiquitous. According to a recent article in The Guardian, more than 150 people died in clashes in one province (Enga) in August 2023 alone.

To help visualize the severity and persistence of current armed conflicts, I made several maps based on the same data found in the Wikipedia article. The first map below is probably the most effective. Rather than shoehorning the data into discrete country-based categories, I placed size-graded stars indicating the 2022 fatality count on the actual location of each conflict, to the extent that that is possible. But it often isn’t, as in the case of the Islamist insurgency in the Maghreb/Sahel, which is listed as occurring in 15 separate countries. At any rate, this map seems more effective at revealing the “clustering” of current conflict than the Wikipedia’s map (posted above). If drug wars are excluded, deadly conflicts in 2022 were concentrated in the “Greater Horn of Africa” (including Yemen), Burma and adjacent parts of South Asia, the northern Middle East proper, central Africa, Nigeria and environs, Afghanistan & Pakistan, and Ukraine.

The map of the duration of current armed conflicts, based on the data in the same Wikipedia table, depicts southern Asia as the area with the most persistent conflicts, followed by Central Africa. The final map shows total fatalities by country in 2021. Whether these maps do a better job of conveying the spatial patterns found in the Wikipedia table than the Wikipedia’s own map is for the reader to decide.

Mapping Recent War Fatalities and the Persistence of Current Armed Conflicts Read More »