Quantifying Urban Diversity: Multiple Spatial Measures of Physical, Social, and Economic Characteristics - Computational Approaches for Urban Environments - page 157

Fig. 7.2 Mix of uses by
street block

when calculating the percentage units within a block group that are residential as
opposed to commercial. Block groups that are geographically larger inherently have
the potential to capture more nonresidential uses. Some authors have addressed this
by utilizing a regularly spaced grid overlaid on the study area and aggregating by
each of the grid squares (Krizek 2003 ). However, by utilizing a grid that is unique
to that particular study, it becomes very difficult, if not impossible, to verify the
study using additional data. Here, by using block groups, there is the possibility
of examining the results of the research against certain types of census data (such
as poverty rates), used as proxies for livability. The utilization of readily available
block groups also facilitates the reproduction of this research within other cities and
by other authors. Other researchers have utilized larger areas of aggregation, such
as Traffic Analysis Zones (TAZs) or census tracts. By way of comparison, there
are 320 TAZs within Washington, D.C. (which are not necessarily contiguous, thus
leaving some areas of the city unstudied), and there are 188 census tracts - far fewer
than the 433 contiguous block groups used in this study.

7.5 Results

After processing the data according to the methodology outlined above, a number of
interesting patterns emerged. Beginning with the first parameter, dwelling density,
the number of dwellings per acre ranged from 0 to 65.97, with a mean of 11.22
and a standard deviation of 10.02. Figure 7.3 shows the distribution of the densities,
as a histogram, and illustrates the rescaled values as an unclassed choropleth map.
Although we do not perform a formal test of similarity to the Poisson distribution