Dot+Density


 * Dot Density Mapping**

Although data frequently are collected for enumeration units and mapped at that level, the resulting maps can be misleading because the distribution of the underlying phenomena often varies within the enumeration units. Using dot maps is a potential solution to this problem.

On a dot map, one dot represents a certain amount of some phenomenon. The dots are placed at locations within the enumeration units where the phenomenon is apt to occur. In this way, the bounds of the symbol are not tied to the enumeration unit (as in choropleth mapping), thus giving the impression that the cartographic representation is more discrete … more precise? … more accurate?
 * The Dot Map**

Dot maps make use of raw totals such as the number of people living in each county. Ancillary information is also necessary for locating dots. For example, if making a dot map of population distribution, we might avoid placing dots in the middle of a lake, on top of a mountain range, or in the desert (presuming that people do not live in these locations).
 * The Data**

Three issues are important to consider when making a dot map: 1) selecting dot size (how large each dot is) and unit value (the raw total represented by each dot) 2) determining regions within which dots should be placed 3) placing dots within selected regions
 * Creating a Dot Map**

Cartographers have argued that the use of very small dots produces a sparse and insignificant distribution map, and the use of very large dots gives an impression of excessive density. Similarly, a small unit value produces a map that can give an unwarranted impression of accuracy, and a unit value that is very large results in a map that “lacks pattern or character”. Generally, it best that the dots in the densest area should just begin to coalesce (merge with one another). In practice, some experimentation is generally required to select an appropriate dot size and unit value.
 * Dot Size and Unit Value**

Generally there are three approaches to dot placement: 1) uniform or geographic placement 2) uniform with a with a random component 3) geographically weighted
 * Placing dots within a region**

Computer mapping programs have been design to automate the dot size and value determination, and placement. But these methods are often undesirable because they often produce unrealistic clustering and gaps in the dot pattern.
 * Automated Methods**


 * General rules for dot density mapping are as follows:**
 * Use an equal-area projection of small to medium scale.
 * Use discrete data.
 * Dot value and dot size should harmonize with the map scale. Dots should just begin to coalesce in the areas of highest density. If possible, two or three dots should be drawn in the enumeration unit that has the least of the mapped quantity. The lower the dot value, the less data are lost. For example, if the dot value is set to 1000, and you have 499 of the mapped quantity in an enumeration unit, then rounding down to 0 means that you “lose” 499 of whatever you are mapping. If the dot value is 500, however, you don’t lose as much of your mapped quantity. Experimentation and compromise is necessary.
 * The dot value must be easily understood. The value 1000 is more appropriate, for example, than 656.
 * Ideally, dots should be placed in the center of gravity of the elements they represent. Because of the random nature of dot placement in desktop mapping, use the smallest enumeration unit possible to achieve a more accurate display of changing density.
 * Avoid accentuating enumeration unit boundaries.
 * Use a legend that shows representative densities.

Back to Thematic Mapping Guides Take me Home