Cartographic issues

Warning View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Limitations of choropleth maps
Ecumene applications at Statistics Canada

The choropleth map is a 'staple' in the cartographic repertoire of thematic mapping. Indeed, over the years the popularity of choropleth maps has increased since this map type is very easy to produce by GIS and cartographic software.

The standard choropleth technique is a method of cartographic representation that employs a distinctive colour or shading that is applied to predefined areal units (Dent 1999, p. 139). For example, the units can be census divisions or census tracts, and the areal symbols cover the entire geographic unit.

Limitations of choropleth maps

An underlying assumption of choropleth mapping is that the data are homogenously or uniformly spread over each geographic unit (MacEachren 1985, p. 42; Dent 1999, p. 141). This assumption is implied by a single colour applied across each unit (Figure 3.1). However, the complete unit or at least a large portion of the unit may be uninhabited, thus producing very misleading spatial patterns (Dorling 1993, p. 170; Crampton 2004, p. 41). In fact, Holloway, Schumacher and Redmond (1997) note that when socioeconomic data are mapped by choropleth techniques:

"…the results often tell us more about the size and shape of the enumeration unit, than about the people actually living and working within them." (Holloway, Schumacher and Redmond 1997, p. 2)

Figure 3.1 Standard choropleth map. Opens a new browser window.

Figure 3.1
Standard choropleth map

Langford and Unwin (1994) endorse this viewpoint by elaborating:

"Although the effects of modifiable boundary location and class interval selection have traditionally been considered the most serious in choropleth mapping, we believe that simple variability in the size and shape of areal units may be of equal importance." (Langford and Unwin 1994, p. 23)

This is illustrated in Figure 3.1. The dark green colour representing the highest data class completely and utterly overpowers the entire map because it covers the larger census divisions in six provinces and two territories. In fact, the largest census division (Baffin located in Nunavut) is about 5,391 times larger in land area than the smallest census division (L'Île-d'Orléans in Quebec).

Ecumene applications at Statistics Canada

The approach taken by Statistics Canada is to use 'generic' ecumenes, such as a population ecumene or an agricultural ecumene. Generic ecumenes confine the statistical distributions to the same areas, and thus render the maps (or map series) more comparable. This goal could not be achieved by using dasymetric maps (Haythornthwaite, Weiss and Heimbecker 1989, p. 192, 194). The dasymetric map is made from the same initial data used for the simple choropleth map. It differs from the choropleth map in that the areas mapped are not bounded solely by predefined census boundaries. Instead, 'natural' boundaries are also taken into account and averages are calculated for each subdivision that is created (Robinson et al. 1995, p. 519, 523).

Population ecumene

Since the 1976 Census, Statistics Canada has applied a population ecumene in its thematic mapping program, primarily for national maps at the census division level1. The use of an ecumene limits the display to only those areas where population is found, resulting in a more accurate depiction of the spatial distribution of data. Dorling (1993, p. 170), Langford and Unwin (1994, p. 23) and others also recommend shading only the inhabited parts of the map.

The methodology for delineating the population ecumene was refined over time. For the 1976 Census, the ecumene was a very rough 'approximation' using 1:7,500,000 population density maps in the fourth edition of the National Atlas of Canada (Natural Resources Canada 1974). The ecumene was generated by delineating those areas that were equal to or greater than one person per square mile (0.4 persons per square kilometre) and then creating a buffer around the areas because the population density maps were based on 1961 data. The population ecumenes for the 1981 and 1986 Censuses were updated to account for population increases, population decreases and minor errors found in the 1976 ecumene.

For the 1991 and 1996 Censuses, the ecumene was defined using the enumeration area as a 'building block', and for the 2001 Census it was defined using a very small geographic unit, the block. The enumeration area and the block are considered to be part of the ecumene if the population density is at least 0.4 persons per square kilometre (about one person per square mile). Section 4 provides further details on the delineation methodology for the 2006 Census.

Figure 3.2 illustrates one of the thematic maps disseminated for the 2006 Census. Note that the population ecumene restricts the data representation to the populated areas. This virtually eliminates the visual perception in Figure 3.1 that larger rural and sparsely populated areas are of greater importance.

Figure 3.2 Choropleth map with 2006 population ecumene (2006 Census of Population data). Opens a new browser window.

Figure 3.2
Choropleth map with 2006 population ecumene (2006 Census of Population data)

The population ecumene developed for census thematic maps is also applied to other Statistics Canada data, such as health data. For example, it was used for the three Mortality Atlases of Canada co-published by Health and Welfare Canada and Statistics Canada in the early 1980s and 1990s (Health and Welfare Canada and Statistics Canada 1980a, 1980b, 1991). In late 2001, maps were added to Health Indicators, an online compilation of data produced jointly by Statistics Canada and the Canadian Institute for Health Information. The maps use the population ecumene at the health region level2 (Figure 3.3).

From time to time, the population ecumene may be applied to maps based on data for the Annual Survey of Manufactures (Figure 3.4).

Figure 3.3 Choropleth map with 2001 population ecumene (2005 health data). Opens a new browser window.

Figure 3.3
Choropleth map with 2001 population ecumene (2005 health data)

Figure 3.4 Choropleth map with 1996 population ecumene (1997 Annual Survey of Manufactures data). Opens a new browser window.

Figure 3.4
Choropleth map with 1996 population ecumene (1997 Annual Survey of Manufactures data)

Agricultural ecumene

Since the 1976 Census, Statistics Canada has also applied an agricultural ecumene for its thematic maps, primarily for national coverage at the census division level. For example, three atlases were produced based on the 1976, 1981 and 1986 Censuses of Agriculture (Statistics Canada 1979, 1984, 1989). Similar to the population ecumene, the agricultural ecumene restricts the data to those areas in which agricultural activity occurs – thereby reducing misinterpretation and visual bias. The atlases for the 1976 and 1986 Censuses also contain dot maps. The ecumene prevents the dots from being randomly spread over entire unit areas or in non-agricultural land.

The methodology for delineating the agricultural ecumene was refined over time. For the earlier censuses, it was created using the enumeration area as the 'building block' (Werschler 1995). By the 2001 Census the ecumene was defined using a small geographic unit, the dissemination area3. It includes all dissemination areas with 'significant' agricultural activity (Figures 3.5 and 3.6). Agricultural indicators, such as the ratio of agricultural land on census farms relative to total land area, and total economic value of agricultural production, are used. Regional variations are also taken into account. The ecumene is generalized for small-scale mapping4 (Statistics Canada 2003, p. 3 to 4; Statistics Canada 2007b).

Figure 3.5 Choropleth map with 2001 agricultural ecumene (2006 Census of Agriculture data). Opens a new browser window.

Figure 3.5
Choropleth map with 2001 agricultural ecumene (2006 Census of Agriculture data)

Figure 3.6 Dot map with 2001 agricultural ecumene (2006 Census of Agriculture data). Opens a new browser window.

Figure 3.6
Dot map with 2001 agricultural ecumene (2006 Census of Agriculture data)


Notes

  1. Census division (CD) is the general term for provincially legislated areas (such as county, municipalité régionale de comté and regional district) or their equivalents. Census divisions are intermediate geographic areas between the province/territory level and the municipality (census subdivision). (Statistics Canada 2007b)
  2. The population ecumene is modified slightly in Quebec for a health region that is in multiple parts (Terres-Cries-de-la-Baie-James).
  3. A dissemination area (DA) is a small, relatively stable geographic unit composed of one or more adjacent dissemination blocks. It is the smallest standard geographic are for which all census data are disseminated. DAs cover all the territory of Canada (Statistics Canada 2007b).
  4. The 2006 agricultural ecumene (to be released in Spring 2008) uses the same delineation criteria as the 2001 ecumene.