Many of the subjects we map in these pages are spatially complex and the choice of geographic scale can have a significant impact on the patterns that appear to the user, especially on the physically limited maps that can be quickly displayed on web pages. Often as not, there are local spatial patterns within metropolitan areas that are overlaid upon regional patterns. To create detailed national maps that allow the user to understand the two levels of patterning, a number of choices can be made.
One of the fundamental challenges for mapping demographics in the United States is the stark differences in the size of geographic areas (block group or county for example) which are relatively small east of the Mississippi and whose size varies substantially in the western states.
For the maps below, we used a single variable – the Demographic Dimensions variable “Affluence” – which conveniently has a mean of zero nationwide at the block group level. We used a consistent number of ranges for display with two color ramps with growing intensity from the mean, but chose an equal number of features in each in order to highlight the effects of different levels of geography.
The first map shows the data at the county level, and while the affluence patterns regionally are clear, they are muted for many urban areas. This occurs because the definition of counties began with a settlement and the surrounding rural area. The income characteristics of the rural areas is often lost at this level of detail:
Contrast this with a block group map of the same variable with the same ‘equal number of features’ per range –
The use of block groups certainly helps in the west, as it becomes clear that affluence is more spatially limited than would appear on the county maps – especially when some of the large counties of Nevada, Arizona, and California are broken down. In the south, we can clearly see that there are affluent cities amongst broad areas of relative poverty.
The regional patterns are clarified substantially when, instead of mapping the actual block group data directly, we map the data for five mile radius areas centered on each block group –
This map much more clearly allows the user to see the broad spatial patterns – especially in the south and in the native lands of the four corners district.
The use of five mile radius areas creates a map, which especially at web viewing physical size and resolution, shows the patterns much more clearly than a simple block group map. The five mile radius summaries are produced using our Snapshot engine, which is able to process the nearly 220,000 radius requests in under 20 minutes on a typical laptop computer. For a minor wait time, this application of a spatial filter can result in maps which clearly convey the forest without entirely losing sight of the individual trees.