Visualizing Hierarchies - Spatial Mathematics: Theory and Practice through Mapping - page 152

7.2 Hierarchies: Census data
Governments have been collecting data, in some cases for hundreds or even
thousands of years, on population, housing, agriculture, economics, and other
demographic variables. The earliest documented censuses date from those
undertaken in 500-499 bc in the Persian Empire (Olmstead, 1948). Sweden's
1749 census is the oldest of the modern census (LSE Library, 2012). Many
countries collect data at several geographic levels, ranging from the country
as a whole, to states and provinces, regions, cities, districts, and individual
city blocks. Some conduct a recurring census every ten years or at other
intervals, and some use the same or similar statistical areas as in the previous
census years. This practice allows analysts to conduct historical comparisons.
The number of questions asked, and the level of detail collected, varies con-
siderably among the statistics agencies conducting the survey. The quality of
data provided in the returns also varies, due to the mobility of the population,
the manner in which the data are collected, the type of government that is
collecting the data, and many other factors. As a result, some sections of the
population may be underrepresented. Some census data are already mapped
for the end user online, while other data are available as tables, and other are
available as GIS-compatible files.
To work effectively with data, it is important to understand the hierarchy
or nested relationship of how the data are collected and represented in
geographic space. For example, land use data are generated from the inter-
pretation of satellite imagery, and its resolution depends on the agency
doing the interpretation and the resolution and type of the original imagery
used. In the case of data from the US Census Bureau, they are collected by
political area and by statistical area. The political areas include American
Indian reservations, cities, counties, and states. Some political areas follow
a standard: Counties nest within states, for example. But much variation
exists beneath that level of geography. In most states, counties include cities.
However, cities also can cross county lines. In Virginia, some cities are not
a part of the county they fall within, but are independent from that county.
Statistical areas, on the other hand, are not political entities but are strictly
for reporting purposes. These include, but are not limited to, census tracts,
block groups, census-designated places, and blocks. Blocks in an urban
area are typically bounded on all sides by streets or some other human-
constructed feature. Blocks in a rural area may be much larger, and may be
bounded by roads, streams, or railroads. Groups of blocks are called block
groups, groups of block groups are called census tracts, and census tracts
nest within counties or county equivalents, such as parishes, independent
cities, or boroughs. The relationships among the various statistical areas are
shown in Figure 7.2 .
In this example, County 9502 is divided into three census tracts. Each census
tract is divided into block groups. Tract 1 is divided into block groups with
numbers in the 100s; Tract 2 into block groups with numbers in the 200s; and