Spatial Thinking in Planning Practice: An Introduction to GIS

29 the location of every individual is known, this method o$en works !ne. If not, the solution is not as simple as it seems. Unfortunately, individual locations are o$en unknown, or they may be con!dential. So$ware like ESRI’s ArcMap, for example, is happy to overlook this shortcoming. Its “Dot Density” option causes point symbols to be positioned randomly within the geographic areas in which the counts were conducted (Figure 4.3). "e size of dots, and number of individuals represented by each dot, are also optional. Random dot placement may be acceptable if the scale of the map is small, so that the areas in which the dots are placed are small. O$en, howev- er, this is not the case. Figure 4.3. A “dot density” map that depicts count data. Source: G. Hatchard. https://www.e-education.psu.edu/ geog482fall2/c3_p17.html An alternative for mapping counts that lack individual locations is to use a single symbol, a circle, square, or some other shape, to represent the total count for each area. ArcMap calls the result of this approach a Proportional Symbol map. When the size of each symbol varies in direct proportion to the data value it represents we have a proportional symbol map (Figure 4.4). In other words, the area of a symbol used to represent the value “1,000,000” is exactly twice as great as a symbol that represents “500,000.” To compensate for the fact that map readers typi- cally underestimate symbol size, some cartographers recommend that symbol sizes be adjusted. ArcMap calls this option “Flannery Compensation” a$er James Flannery, a research cartographer who conducted psychophysical studies of map symbol perception in the 1950s, 60s, and 70s. A variant on the Proportional Symbol approach is the Graduated Symbol map type, in which di#erent symbol sizes represent categories of data values rather than unique values. In both of these map types, symbols are usually placed at the mean locations, or centroids, of the areas they represent. Chapter 4: Mapping People with Census Data