Tract-to-CCA Crosswalk — xwalk_tract2cca • cmapgeo

This table contains a set of factors to apportion Census tract-level data among Chicago Community Areas (CCAs). Separate factors are provided for apportioning housing unit, household, and population attributes. All factors were determined by calculating the percentage of a tract's housing units, households and population that were located in each of its component blocks, according to the 2020 Decennial Census, and then assigning each block to a CCA (based on the location of the block's centroid point). Use xwalk_tract2cca for data from the 2020 decennial census or the American Community Survey (ACS) from 2020 onward. For data from the 2010 decennial census or ACS from 2010 through 2019, use xwalk_tract2cca_2010.

xwalk_tract2cca

xwalk_tract2cca_2010

Format

xwalk_tract2cca is a tibble with 818 rows and 6 variables:

geoid_tract: Unique 11-digit tract ID, assigned by the Census Bureau. Corresponds to tract_sf. Character.
cca_num: Numeric CCA ID, as assigned by the City of Chicago. Corresponds to cca_sf. Integer.
hu_pct: Proportion of the tract's housing units (occupied or vacant) located in the specified CCA. Multiply this by a tract-level measure of a housing attribute (e.g. vacant homes) to estimate the CCA's portion. Double.
hh_pct: Proportion of the tract's households (i.e. occupied housing units) living in the specified CCA. Multiply this by a tract-level measure of a household attribute (e.g. car-free households) to estimate the CCA's portion. Double.
pop_pct: Proportion of the tract's total population (including group quarters) living in the specified CCA. Multiply this by a tract-level measure of a population attribute (e.g. race/ethnicity) to estimate the CCA's portion. Double.
emp_pct: Proportion of the tract's total jobs located in the specified CCA. Multiply this by a tract-level measure of an employment attribute (e.g. retail jobs) to estimate the CCA's portion. Not available in xwalk_tract2cca_2010. Double.

xwalk_tract2cca_2010 is a tibble with 805 rows and 5 variables (no emp_pct).

Details

Generally speaking, tract boundaries align neatly with CCA boundaries as they tend to follow similar features (e.g. rivers, major roads, rail lines) but there are cases where the jobs, population, households and/or housing units in a tract are split across multiple CCAs, or else are partially within the City of Chicago and partially outside of it. For that reason, it is not appropriate to use a one-to-one tract-to-CCA assignment to apportion Census data among CCAs, and this crosswalk should be used instead.

To use this crosswalk effectively, Census data should be joined to it (not vice versa, since tract IDs appear multiple times in this table). Once the data is joined, it should be multiplied by the appropriate factor (depending whether the data of interest is measured at the housing unit, household, person or job level), and then the result should be summed by CCA. If calculating rates, this should only be done after the counts have been summed to CCA. The resulting table can then be joined to cca_sf for mapping, if desired.

If your data is also available at the block group level, it is recommended that you use that with xwalk_blockgroup2cca instead of the tract-level allocation.

Examples

suppressPackageStartupMessages(library(dplyr))

# View the tracts with population split between multiple CCAs
filter(xwalk_tract2cca, pop_pct < 1)
#> # A tibble: 34 × 6
#>    geoid_tract cca_num  hu_pct  hh_pct pop_pct emp_pct
#>    <chr>         <int>   <dbl>   <dbl>   <dbl>   <dbl>
#>  1 17031081403      32 0       0       0       0.205  
#>  2 17031161100      15 0       0       0       0.193  
#>  3 17031430400      69 0       0       0       0.0279 
#>  4 17031520500      52 0.997   0.997   0.994   1      
#>  5 17031520500      55 0.00254 0.00264 0.00588 0      
#>  6 17031520600      52 0.993   0.993   0.981   1      
#>  7 17031520600      55 0.00712 0.00738 0.0194  0      
#>  8 17031550200      54 0       0       0       0.008  
#>  9 17031580300      30 0       0       0       0.218  
#> 10 17031590700      60 0       0       0       0.00259
#> # ℹ 24 more rows

# Estimate CCA-level population density from tract-level Census data
df_tract <- tidycensus::get_decennial(
  geography = "tract", variables = c("P1_001N"),
  year = 2020, state = "IL", county = c("031", "043"), output = "wide"
) %>%
  suppressMessages() %>%  # Hide tidycensus messages
  select(geoid_tract = GEOID, pop = P1_001N)

df_cca <- xwalk_tract2cca %>%
  left_join(df_tract, by = "geoid_tract") %>%
  mutate(pop = pop * pop_pct) %>%
  group_by(cca_num) %>%
  summarize(pop = sum(pop))
df_cca
#> # A tibble: 77 × 2
#>    cca_num    pop
#>      <int>  <dbl>
#>  1       1  55628
#>  2       2  77122
#>  3       3  57182
#>  4       4  40494
#>  5       5  35114
#>  6       6 103050
#>  7       7  70492
#>  8       8 105481
#>  9       9  11525
#> 10      10  38303
#> # ℹ 67 more rows

# Join to cca_sf for mapping
library(ggplot2)
cca_sf %>%
  left_join(df_cca, by = "cca_num") %>%
  ggplot() +
    geom_sf(aes(fill = pop / sqmi), lwd = 0.1) +
    scale_fill_viridis_c(direction = -1) +
    theme_void()