Choropleth Maps

Normalization

Choropleth maps involve mapping data by areal unit, for example by country, state, or county. Because of the disparities in size that exist for all types of areal units, mapping raw totals will often present a skewed representation of the data. To combat this problem, choropleth maps are frequently normalized by a common attribute, such as area or population. Once normalized, the maps are often referred to as value-by-area or value-by-population maps. By normalizing the raw data, the map is not dominated by the largest unit but rather uses the ratio to highlight the ratios of interest.

Representation

Choropleth maps occur in two distinct types: classed and unclassed. A classed choropleth map presents the data of interest by classifying the values into a distinct number of individual colored classes (Figure 1). Each class is presented as a range of possible values, and the units within each range are represented accordingly.

In contrast, an unclassed choropleth map presents the value of each unit in a distinct color that varies according to the rank of the value with respect to all the other values (Figure 2). Traditionally, both types of choropleth map represent low values with light color and use dark colors to represent greater values.

Figure 1 Classed choropleth map of population density for contiguous United States

Source: Adapted by author from U.S. Census Bureau data. Available from http://www.census.gov/population/www/popdata.html.

Figure 2 Unclassed choropleth map of population density for contiguous United States

Source: Adapted by author from U.S. Census Bureau data. Available from http://www.census.gov/population/www/popdata.html.

Data Classification

When a classed choropleth map is created, the classification type must be decided based on the data. One of the most common classification schemes used is the Natural Breaks method. This method involves plotting all data values on a graph and then identifying the largest gaps between the data clusters. Most mapping software will perform this analysis automatically. The other common classification schemes follow more traditional statistical measures, such as quantile classification, defined intervals, or mean/standard deviation classification (most commonly used for diverging data with a meaningful middle). Often, these classification schemes are used as a basis for starting the classification but are modified to accommodate meaningful data values that may fail to be captured by an automated classification method.

Symbolization

Choropleth maps are symbolized in one of three ways, sequentially, diverging, or qualitatively. Pattern is largely limited to qualitative symbolization, though it can be used in combination with color to extend a sequential classification. Black-and-white color tends to restrict the map to a sequential representation—unless combined with pattern. Full color allows for all the options (Figure 3). Sequential symbolization follows a light to [Page 408]dark path either through a single hue or from one hue to a second, sometimes through a third. Diverging symbolization places two dark colors at the poles of the symbolization and lightens each toward the middle, where a white or near-white color represents the meaningful middle value. Qualitative colors use differing hues to represent the different values, often with nonnumeric data.

...