Fig 1.
Illustration of the Racial Landscape (RL) method.
A-D: Data transformation from a census block (A) to the RL grid (D). In this illustration, all cells in (D) have the same densities, but in general, densities vary. Different colors reflect different racial sub-populations. E: An example of the RL image showing the racial geography of parts of Washington, DC. Each cell has one of six colors corresponding to six contributing racial sub-populations, but same-race cells may differ in shades, depicting population density—the darker the shade, the larger the local population density. F: The RL image of the area enclosed in the purple rectangle. At this magnification, the details of population distribution begin to emerge. Uninhabited areas are shown in white.
Fig 2.
Calculating a metric of racial segregation using the RL method.
A: RL grid with each cell having a given race and storing the value of local population density. B: An exposure matrix calculated from RL grid layers. The exposure matrix is a co-occurrence matrix weighted by an average of local population densities in neighboring cells. C: An example of calculating an entry to the exposure matrix using the red-green (Black-Asian) cells. D-F: Examples of RL image visualizations of three 12km × 12km areas in Washington DC and values of NH, MI, and NMI calculated from them.
Fig 3.
Outline of our procedure to calculate the National Racial Geography Dataset 2020.
Numbers are pointers used in the main text.
Table 1.
The list of layers in the NRGD2020.
Fig 4.
An example of mapping racial diversity at different length scales.
A: The RL image of the Los Angeles area serves as the reference map. B-D: Racial diversity maps at length scales of 0.75 km, 3 km, and 12 km, respectively, overlaying the RL image. E-G: Examples of three locations in the Los Angeles area exhibiting low, medium, and high racial diversity, respectively. The pointers indicate the mapping of these locations on the diversity maps.
Fig 5.
An example of mapping racial segregation at different length scales.
A: The RL image of the rural area in Mississippi (referred to as the “Choctaw site”) serves as the reference map. B-D: Racial segregation maps at length scales of 3 km, 6 km, and 12 km, respectively, overlying the RL image. E-G: Segregation maps at the three length scales for a single 12 km × 12 km sub-tile marked by a star symbol on panel (D).
Fig 6.
RL image of Atlanta MSA overlaid by the boundaries of the four zones.
The red boundary outlines the principal city, black boundaries outline three types of suburbs, inner-ring (A), outlying (B), and fringe (C). The table lists values of diversity and segregation metrics for the zones.
Fig 7.
Changes in segregation level within a scale in six MSAs.
A: Segregation profiles for six MSAs, see main text for details. B: A sequence of maps showcasing the Jackson MSA, RL image followed by three NRGD2020 precalculated segregation maps at length scales of 3, 9, and 18 km, respectively. C: The same content as (B) but for the Chicago MSA.
Table 2.
Comparison of traditional approach and RL method in analyzing and visualizing racial geography.