The New York Times’ Flubbed China Cartograms

NY Times China Population CartogramAn interesting story in today’s (April 9) New York Times—“Hello, Cambodia: Wary of Events in China, Foreign Investors Head to the South”—is illustrated in the print edition with two striking cartograms of eastern Asia, one of which shows population and the other economic output. The cartogram legends claims that “countries and Chinese provinces are sized according to population” and, respectively to “economic output.” Actually, they are not. On the population cartogram, for example, compare the sizes of Hong Kong and Taiwan with that of Thailand. Is Thailand shown as almost ten times larger than Hong Kong and almost three times the size of Taiwan, as an accurate depiction would have it? Hardly.


NY Times China Economic CartogramThe real problem with the maps, however, is the claim that Chinese provinces are also sized according to these metrics. In actuality, it appears that no efforts were made to depict China’s first-order internal divisions (which include autonomous regions and direct-controlled municipalities in addition to standard provinces) in the manner of a cartogram. If this had been done, China would not retain its familiar shape, as can immediately be seen on an actual population cartogram of the country, produced by Worldmapper. On an economic cartogram, the shape distortion would be even more pronounced, as production is concentrated in the coastal provinces. As the Economist map shows, the GDP of the Tibetan Autonomous Region is roughly equivalent to that of Malta.

WorldMapperChinaPopulationCartogramThe New York Times cartograms also seemingly imply that Hong Kong is an independent country, rather than a “special administrative region” of China.




Non-State-Based Atlas Preface, Part II

Wikipedia map of nominal per capita GDP, 2007Maps and text from the forthcoming non-state-based (or “demic”) atlas will begin appearing in GeoCurrents next week. This week, the blog is presenting the work’s preface.

As noted in the previous post, countries are incomparable units, due to their vast variation in scale. Yet in tables and charts, Nauru, with ten thousand inhabitants living on eight square miles (twenty-one square kilometers), counts the same as China, with 1.3 billion inhabitants living on 3.7 million square miles (9.6 million square kilometers). Cartographic depictions are inherently less distorting, as they are spatially scaled. On most maps, small countries are appropriately small, although only those using equal-area projections maintain strict size proportionality. Miniscule states such as Nauru and Monaco thus tend to vanish from view. Unfortunately, mapmakers seem to be increasingly depicting micro-states with large circles, again privileging diplomatic pretense over geographical reality. Curiously, the Wikipedia GDP map posted here balloons Europe’s feudal remnants by several orders of magnitude, yet does not provide any information about them.

Although spatial imbalances are minimized on most maps, demographic disparities remain concealed. True, tiny countries usually have tiny populations, just as large ones tend to have many inhabitants, but the correlation is not strong. Confined to 268 square miles (694 square kilometers), Singapore’s five million inhabitants have little presence on standard maps, whereas Mongolia’s 2.8 million residents figure large in their 603,909 square-mile homeland (1,564,115 square kilometers). Unless one has a good sense of the distribution of human settlement across the world, maps of socio-economic development tend to mislead. On a world map of per capita GDP, wealthy Australia and Canada far overshadow India, even though India’s population is twenty-five times greater than that of the other two countries combined. In Africa, areas of relatively high per capita GDP are not as significant as they appear. Botswana, Namibia, and Gabon are sizable countries that boast elevated economic figures, but together they contain five and a half million people, roughly half the population of the destitute metropolis of Kinshasa in the Democratic Republic of Congo.

Cartogram of World Population Mapmakers have devised a number of innovative techniques for depicting data in a more proportional manner. In demographic cartograms, the areas of geographically defined entities are expanded or reduced according to their populations. GDP figures are also well suited to cartogramic treatment, as both total economic output and output per capita can be depicted by a combination of size and color. The resulting image posted below nicely illustrates the concentration of economic power in a few parts of the world, as well as its near absence over a large swath of central Africa.

Cartogram of GDPBut for all of their powers of presentation, global cartograms conceal as much as they reveal. Areas with low numerical values in regard to whatever feature is being mapped tend to shrink into invisibility. On the example posted here, the Democratic Republic of Congo, a massive country of more than seventy million, can hardly be discerned. More to the point, cartograms of global development are almost always structured around sovereign states, and thus treat massive and highly differentiated countries as uniform entities. Note that this cartogram portrays Taiwan as if it really were part of China, rather than as the de facto sovereign state that it is; as a result, Taiwan appears smaller than Sri Lanka, even though its economy is almost an order of magnitude larger. (Why the cartographer depicts uninhabited Antarctica as having an economy roughly the size of Canada’s at a per capita level similar to that of India is anyone’s guess.)

Map of GDP densityAnother method of addressing disparities of scale when mapping economic production is encountered in the GDP density map, analyzed earlier in this blog by Andrew Lindford. Portraying “GDP per square kilometer” is an intriguing idea, but the effort fails, as the cartographer treats countries as undifferentiated wholes in regard to economic but not demographic data. The result is merely a map of settlement density in which the populations of wealthy countries are weighed more heavily than those of poor countries. Treating countries as economic uniformities again results in nonsense. Note how China’s Sichuan basin is shown as more economically productive per unit of area than its lower Yangtze region; observe how India’s impoverished lower Ganges Valley is depicted as more “economically dense” than its bustling greater Mumbai area.

The root problem is clear: basic patterns of social and economic development do not necessarily track the contours of political geography. To be sure, the immediate gap across national boundaries can be profound. When one moves from the U.S. state of New Mexico to the Mexican state of Chihuahua, per capita GDP drops three-fold, from $36,000 to $12,300. Yet the economic gaps within Mexico are more substantial than the chasm along its northern border. The southern Mexican state of Chiapas posts a per capita GDP figure (in PPP) of only $3,700, whereas Nuevo León in the north comes in at $16,300 and the central Federal District reaches $23,000, the latter figure not out of line with Mississippi’s $32,000. Similarly discrepant patterns are apparent across much the world. In socio-economic terms, southern Brazil is much more similar to Argentina than it is to northeastern Brazil, just as northern Italy is more akin to Switzerland than to southern Italy. A state-based system of comparison obscures all such internal differentiation.

If we are to devise more appropriate methods of portraying global social and economic disparities, we must move beyond the default framework of sovereign states and their dependent territories. This is not to claim that the governments and policies of independent countries are insignificant in determining which areas of the world are more prosperous or healthier than others. Nor is it to argue that conventional country-based maps of socio-economic development should be jettisoned, as they do serve a significant purpose. It is rather merely to insist that sovereign states are not the all-important, all-purpose units of global geography that they ubiquitously taken to be. As a result, the standard state-based map should be complemented by other modes of presentation. The current project is a first step in that direction.

Microstates in Cartograms

Microstates such as Lichtenstein or Nauru are too small to be seen on most world maps, and even a country as large as Luxembourg is usually difficult to discern. In conventional cartography, the size of an area depicted on the map is roughly proportional to its actual size, consigning tiny countries to invisibility. But not all maps are based on area. In cartograms, other variables substitute, distorting the size (and shape) of geographical entities to show their relative rankings according to any number of measurements. In a world population cartogram (first map above), the size of each country is proportional to its number of inhabitants. In the second map, GDP determines size.

In world population cartograms, microstates generally remain imperceptible. Singapore might be considered a major exception. Although only a quarter the area of Luxembourg, Singapore, with five million inhabitants, is easily seen here. Singapore may be a micro-country, but given its sizable population, it is not a microstate. Cartograms can thus effectively portray important variations that escape conventional cartography, such as the importance of Singapore. To see what else can be done with cartograms, visit the impressive collection on the Worldmapper website.

Worldmapper does not include truly tiny sovereign states, but it does map the world’s merely small countries, including Luxembourg. It is instructive to see how Luxembourg appears on its maps, given its economic strength. On the GDP cartogram posted above (made by a Worldmapper collaborator), Luxembourg is small but easily found. But on Worldmapper’s depiction of “finance and insurance service exports” – surely the most bizarre image of the world I have ever seen – it looms large as a veritable macro-state.

Useful as it is, the Worldmapper project is not without problems. Much of the data employed is seriously out of date; its internet penetration map, based on 2002 information, is of historical interest only. More inherent difficulties stem – yet again – from the use of a state-based framework. Cartograms are invariably political maps of a sort, as numbers must be gathered for clearly demarcated areas, which are almost always countries or subdivisions thereof. Political maps, however, are not the proper vehicles for all cartographic purposes: the Worldmapper portrayal of “species existing in zoos only” is particularly unenlightening. Worldmapper also scales dependencies in accordance with the countries that hold their sovereignty. As a result, the Falklands Islands (Malvinas) figure as part of the United Kingdom, and are thus enlarged on all of the maps posted above, grotesquely so in the financial export map (note that even uninhabited South Georgia is depicted as a huge island here). Yet oddly enough, Worldmapper colors both the Falklands Islands and South Georgia as if they were Latin American countries.

Worldmapper’s motto, “the world as you have never seen it before,” is apt. But I find that different people react quite differently to its maps. Some find them fascinating, others arresting, and others disturbing. For teaching purposes, I find cartograms most effective, but only if used sparingly.