Bayesian Improved Surname Geocoding (BISG)
This tutorial demonstrates how to perform Bayesian Improved Surname Geocoding when the race/ethncity of individuals are unknown within a dataset.
What is Bayesian Improved Surname Geocoding?
Bayesian Improved Surname Geocoding (BISG) is a method that applies Bayes’ Rule to predict the race/ethnicity of an individual using the individual’s surname and geocoded location [Elliott et. al 2008, Elliot et al. 2009, Imai and Khanna 2016].
Specifically, BISG first calculates the prior probability of individual i being of a ceratin racial group r given their surname s or \[Pr(R_i=r|S_i=s)\]. The prior probability created from the surname is then updated with the probability of the individual i living in a geographic location g belonging to a racial group r, or \[Pr(G_i=g|R_i=r)\]). The following equation describes how BISG calculates race/ethnicity of individuals using Bayes Theorem, given the surname and geographic location, and specifically when race/ethncicty is unknown :
\[Pr(R_i=r|S_i=s, G_i=g)=\frac{Pr(G_i= g|R_i =r)Pr(R_i =r |S_i= s)}{\sum_{i=1}^n Pr(G_i= g|R_i =r)Pr(R_i =r |S_i= s)}\]
In R, the wru package titled, WRU: Who Are You
performs BISG. This vignette will walk you through how to prepare your
geocoded voter file for performing BISG by stepping you through the
process of cleaning your voter file, prepping voter data for running the
BISG, and finally, performing BISG to obtain racial/ethnic probailities
of individuals in a voter file.
Performing BISG on your data
We will perform BISG using the previous Gwinnett and Fulton county
voter registration data called ga_geo.csv that was geocoded
in the eiCompare: Geocoding vignette.
The first step in performing BISG is to geocode your voter file addresses. For information on geocoding, visit the Geocoding Vignette.
Let’s begin by loading your geocoded voter data into R/RStudio.
Step 1: Load R libraries/packages, voter file, and census data
Load the R packages needed to perform BISG. If you have not already downloaded the following packages, please install these packages.
# Load libraries
suppressPackageStartupMessages({
library(eiCompare)
library(stringr)
library(sf)
library(wru)
library(tidyr)
library(ggplot2)
library(dplyr)
})Load in census data, the shape file and geocoded voter registration data with latitude and longitude coordinates.
# Load Georgia census data
data(georgia_census)We will use the data(gwin_fulton_shape) to load the
shape file. The shape file includes FIPS code information for Gwinnett
and Fulton counties and the associated multipolygon shape geometries
indicated by the geometry column.
# Shape file for Gwinnett and Fulton counties
data(gwin_fulton_shape)Loading a shapefile using the tigris package
(optional)
The shapefile can also be loaded using the tigris package. The tigris package uses the US Census Bureau’s Geocoding API which is publicly available so no API key is needed. With the tigris package, you can load your census data according to a geographic level (i.e. counties, cities, tracts, blocks, etc.) There is additional code below that you can use if wanting to load your shape file using tigris. Remember to remove the # in order to use the code.
# install.packages("tigris")
# library(tigris)
# gwin_fulton_shape <- blocks(state = "GA", county = c("Gwinnett", "Fulton"))Load geocoded voter file.
# Load geocoded voter registration file
data(ga_geo)Obtain the first six rows of the voter file to check that the file has downloaded properly.
# Check the first six rows of the voter file
head(ga_geo, 6)
#> # A tibble: 6 × 25
#> county…¹ count…² regis…³ voter…⁴ last_…⁵ first…⁶ str_num str_n…⁷ str_s…⁸ city
#> <dbl> <chr> <dbl> <chr> <chr> <chr> <dbl> <chr> <chr> <chr>
#> 1 60 Fulton 1 A LOCKLER GABRIE… 1084 Howell… NW Atla…
#> 2 60 Fulton 2 A RADLEY OLIVIA 7305 Villag… <NA> Fair…
#> 3 60 Fulton 3 A BOORSE KEISHA 6200 Bakers… SW Atla…
#> 4 67 Gwinne… 12 A MAZ SAVANN… 1359 Beaver… <NA> Norc…
#> 5 67 Gwinne… 13 A GAULE NATASH… 2961 Lenora… <NA> Snel…
#> 6 67 Gwinne… 15 A MCMELL… ISMAEL 4305 Paxton… <NA> Lilb…
#> # … with 15 more variables: state <chr>, zipcode <dbl>, street_address <chr>,
#> # final_address <chr>, cxy_address <chr>, cxy_status <chr>,
#> # cxy_quality <chr>, cxy_matched_address <chr>, cxy_tiger_line_id <dbl>,
#> # cxy_tiger_side <chr>, STATEFP10 <chr>, COUNTYFP10 <chr>, TRACTCE10 <chr>,
#> # BLOCKCE10 <chr>, geometry <chr>, and abbreviated variable names
#> # ¹county_code, ²county_name, ³registration_number, ⁴voter_status,
#> # ⁵last_name, ⁶first_name, ⁷str_name, ⁸str_suffix
#> # ℹ Use `colnames()` to see all variable namesView the column names of the voter file. Some of these columns will be used along the journey to performing BISG.
# Find out names of columns in voter file
names(ga_geo)
#> [1] "county_code" "county_name" "registration_number"
#> [4] "voter_status" "last_name" "first_name"
#> [7] "str_num" "str_name" "str_suffix"
#> [10] "city" "state" "zipcode"
#> [13] "street_address" "final_address" "cxy_address"
#> [16] "cxy_status" "cxy_quality" "cxy_matched_address"
#> [19] "cxy_tiger_line_id" "cxy_tiger_side" "STATEFP10"
#> [22] "COUNTYFP10" "TRACTCE10" "BLOCKCE10"
#> [25] "geometry"Check the dimensions (the number of rows and columns) of the voter file.
# Get the dimensions of the voter file
dim(ga_geo)
#> [1] 12 25There are 12 voters (or observations) and 25 columns in the voter file.
Convert geometry column name into two columns for latitude and longitude points.
ga_geo <- ga_geo %>%
tidyr::extract(geometry, c("lon", "lat"), "\\((.*), (.*)\\)", convert = TRUE)Step 2: De-duplicate the voter file.
The next step involves removing duplicate voter IDs from the voter
file, using the dedupe_voter_file function.
# Remove duplicate voter IDs (the unique identifier for each voter)
voter_file_dedupe <- dedupe_voter_file(voter_file = ga_geo, voter_id = "registration_number")There are no duplicate voter IDs in the dataset.
Step 3: Perform BISG and obtain the predicted race/ethnicity of each voter.
# Convert the voter_shaped_merged file into a data frame for performing BISG.
voter_file_complete <- as.data.frame(voter_file_dedupe)
class(voter_file_complete)
#> [1] "data.frame"Note that wru requires an internet connection to pull in
supplemental data. If the connection cannot be made,
wru_predict_race_wrapper will return NULL.
georgia_census$GA$year <- 2010
# Perform BISG
bisg_df <- eiCompare::wru_predict_race_wrapper(
voter_file = voter_file_complete,
census_data = georgia_census,
voter_id = "registration_number",
surname = "last_name",
state = "GA",
county = "COUNTYFP10",
tract = "TRACTCE10",
block = "BLOCKCE10",
census_geo = "block",
use_surname = TRUE,
surname_only = FALSE,
surname_year = 2010,
use_age = FALSE,
use_sex = FALSE,
return_surname_flag = TRUE,
return_geocode_flag = TRUE,
verbose = TRUE
)
#> Matching surnames.
#> Performing BISG to obtain race probabilities.
#> Proceeding with last name predictions...
#> Proceeding with Census geographic data at block level...
#> Using Census geographic data from provided census.data object...
#> State 1 of 1: GA
#> ℹ All local files already up-to-date!
#> BISG complete.# Check BISG dataframe
head(bisg_df)
#> county_code county_name registration_number voter_status last_name first_name
#> 1 60 Fulton 1 A LOCKLER GABRIELLA
#> 2 60 Fulton 2 A RADLEY OLIVIA
#> 3 60 Fulton 3 A BOORSE KEISHA
#> 4 67 Gwinnett 12 A MAZ SAVANNAH
#> 5 67 Gwinnett 13 A GAULE NATASHIA
#> 6 67 Gwinnett 15 A MCMELLEN ISMAEL
#> str_num str_name str_suffix city state zipcode
#> 1 1084 Howell Mill Rd NW NW Atlanta GA 30318
#> 2 7305 Village Center Blvd <NA> Fairburn GA 30213
#> 3 6200 Bakers Ferry Rd SW SW Atlanta GA 30331
#> 4 1359 Beaver Ruin Road <NA> Norcross GA 30093
#> 5 2961 Lenora Church Rd <NA> Snellville GA 30078
#> 6 4305 Paxton Ln <NA> Lilburn GA 30047
#> street_address final_address
#> 1 1084 Howell Mill Rd NW NW 1084 Howell Mill Rd NW NW,Atlanta,GA,30318
#> 2 7305 Village Center Blvd 7305 Village Center Blvd,Fairburn,GA,30213
#> 3 6200 Bakers Ferry Rd SW SW 6200 Bakers Ferry Rd SW SW,Atlanta,GA,30331
#> 4 1359 Beaver Ruin Road 1359 Beaver Ruin Road,Norcross,GA,30093
#> 5 2961 Lenora Church Rd 2961 Lenora Church Rd,Snellville,GA,30078
#> 6 4305 Paxton Ln 4305 Paxton Ln,Lilburn,GA,30047
#> cxy_address cxy_status cxy_quality
#> 1 1084 Howell Mill Rd NW NW, Atlanta, GA, 30318 Match Non_Exact
#> 2 7305 Village Center Blvd , Fairburn, GA, 30213 Match Exact
#> 3 6200 Bakers Ferry Rd SW SW, Atlanta, GA, 30331 Match Exact
#> 4 1359 Beaver Ruin Road , Norcross, GA, 30093 Match Exact
#> 5 2961 Lenora Church Rd , Snellville, GA, 30078 Match Exact
#> 6 4305 Paxton Ln , Lilburn, GA, 30047 Match Exact
#> cxy_matched_address cxy_tiger_line_id
#> 1 1084 HOWELL MILL RD, ATLANTA, GA, 30318 17341994
#> 2 7305 VILLAGE CENTER BLVD, FAIRBURN, GA, 30213 650118829
#> 3 6200 BAKERS FERRY RD SW, ATLANTA, GA, 30331 17378238
#> 4 1359 BEAVER RUIN RD, NORCROSS, GA, 30093 638089058
#> 5 2961 LENORA CHURCH RD, SNELLVILLE, GA, 30078 647930651
#> 6 4305 PAXTON LN, LILBURN, GA, 30047 88948522
#> cxy_tiger_side STATEFP10 COUNTYFP10 TRACTCE10 BLOCKCE10 lon lat
#> 1 L 13 121 000600 1064 -84.41194 33.78431
#> 2 L 13 121 010514 3038 -84.59827 33.55365
#> 3 L 13 121 010303 2019 -84.57126 33.72542
#> 4 L 13 135 050424 1011 -84.14037 33.92892
#> 5 R 13 135 050719 1008 -84.00938 33.83915
#> 6 R 13 135 050714 2027 -84.07677 33.83707
#> matched_surname matched_geocode pred.whi pred.bla pred.his
#> 1 TRUE TRUE 8.236113e-07 2.200425e-08 2.129911e-08
#> 2 TRUE TRUE 1.442265e-01 8.279974e-01 7.321964e-03
#> 3 TRUE TRUE 9.699416e-01 0.000000e+00 2.532548e-02
#> 4 TRUE TRUE 5.214638e-02 5.520133e-02 7.600797e-01
#> 5 TRUE TRUE 6.707808e-01 1.917867e-01 1.374325e-01
#> 6 TRUE TRUE 7.297028e-01 1.509212e-01 5.958378e-02
#> pred.asi pred.oth
#> 1 0.0000000000 3.836384e-07
#> 2 0.0007539897 1.970021e-02
#> 3 0.0047329064 0.000000e+00
#> 4 0.1325725492 0.000000e+00
#> 5 0.0000000000 0.000000e+00
#> 6 0.0284345396 3.135771e-02Summarizing BISG output
summary(bisg_df)
#> county_code county_name registration_number voter_status
#> Min. :60.00 Length:12 Min. : 1.00 Length:12
#> 1st Qu.:65.25 Class :character 1st Qu.: 9.00 Class :character
#> Median :67.00 Mode :character Median :14.00 Mode :character
#> Mean :65.25 Mean :12.25
#> 3rd Qu.:67.00 3rd Qu.:17.25
#> Max. :67.00 Max. :20.00
#> last_name first_name str_num str_name
#> Length:12 Length:12 Min. : 287 Length:12
#> Class :character Class :character 1st Qu.:1394 Class :character
#> Mode :character Mode :character Median :3498 Mode :character
#> Mean :3306
#> 3rd Qu.:4162
#> Max. :7305
#> str_suffix city state zipcode
#> Length:12 Length:12 Length:12 Min. :30045
#> Class :character Class :character Class :character 1st Qu.:30078
#> Mode :character Mode :character Mode :character Median :30094
#> Mean :30167
#> 3rd Qu.:30239
#> Max. :30518
#> street_address final_address cxy_address cxy_status
#> Length:12 Length:12 Length:12 Length:12
#> Class :character Class :character Class :character Class :character
#> Mode :character Mode :character Mode :character Mode :character
#>
#>
#>
#> cxy_quality cxy_matched_address cxy_tiger_line_id cxy_tiger_side
#> Length:12 Length:12 Min. : 17341994 Length:12
#> Class :character Class :character 1st Qu.: 88908516 Class :character
#> Mode :character Mode :character Median :637257860 Mode :character
#> Mean :399987381
#> 3rd Qu.:641834456
#> Max. :650118829
#> STATEFP10 COUNTYFP10 TRACTCE10 BLOCKCE10
#> Length:12 Length:12 Length:12 Length:12
#> Class :character Class :character Class :character Class :character
#> Mode :character Mode :character Mode :character Mode :character
#>
#>
#>
#> lon lat matched_surname matched_geocode
#> Min. :-84.60 Min. :33.55 Mode:logical Mode:logical
#> 1st Qu.:-84.23 1st Qu.:33.82 TRUE:12 TRUE:12
#> Median :-84.14 Median :33.89
#> Mean :-84.19 Mean :33.88
#> 3rd Qu.:-84.04 3rd Qu.:33.97
#> Max. :-83.97 Max. :34.09
#> pred.whi pred.bla pred.his pred.asi
#> Min. :0.0000008 Min. :0.000000 Min. :0.000000 Min. :0.0000000
#> 1st Qu.:0.0288208 1st Qu.:0.008116 1st Qu.:0.007098 1st Qu.:0.0005655
#> Median :0.3390319 Median :0.103061 Median :0.026733 Median :0.0034098
#> Mean :0.4106619 Mean :0.232251 Mean :0.169077 Mean :0.0896576
#> 3rd Qu.:0.7395151 3rd Qu.:0.241425 3rd Qu.:0.079046 3rd Qu.:0.0164522
#> Max. :1.0000000 Max. :0.950896 Max. :0.943675 Max. :0.8884684
#> pred.oth
#> Min. :0.000000
#> 1st Qu.:0.000000
#> Median :0.007979
#> Mean :0.015019
#> 3rd Qu.:0.022615
#> Max. :0.064206Look at BISG race predictions by county
# Obtain aggregate values for the BISG results by county
bisg_agg <- precinct_agg_combine(
voter_file = bisg_df,
group_col = "COUNTYFP10",
race_cols = c("pred.whi", "pred.bla", "pred.his", "pred.asi", "pred.oth"),
true_race_col = "race",
include_total = FALSE
)
# Table with BISG race predictions by county
head(bisg_agg)
#> # A tibble: 2 × 6
#> COUNTYFP10 pred.whi_prop pred.bla_prop pred.his_prop pred.asi_prop pred.oth_…¹
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 121 0.371 0.276 0.0109 0.00183 0.00657
#> 2 135 0.424 0.218 0.222 0.119 0.0178
#> # … with abbreviated variable name ¹pred.oth_propBarplot of BISG results
bisg_bar <- bisg_agg %>%
tidyr::gather("Type", "Value", -COUNTYFP10) %>%
ggplot(aes(COUNTYFP10, Value, fill = Type)) +
geom_bar(position = "dodge", stat = "identity") +
labs(title = "BISG Predictions for Fulton and Gwinnett Counties", y = "Proportion", x = "Counties") +
theme_bw()
bisg_bar + scale_color_discrete(name = "Race/Ethnicity Proportions")
Choropleth Map
Finally, we will map the BISG data onto choropleth maps.
bisg_dfsub <- bisg_df %>%
dplyr::select(BLOCKCE10, pred.whi, pred.bla, pred.his, pred.asi, pred.oth)
bisg_dfsub
#> BLOCKCE10 pred.whi pred.bla pred.his pred.asi pred.oth
#> 1 1064 8.236113e-07 2.200425e-08 2.129911e-08 0.0000000000 3.836384e-07
#> 2 3038 1.442265e-01 8.279974e-01 7.321964e-03 0.0007539897 1.970021e-02
#> 3 2019 9.699416e-01 0.000000e+00 2.532548e-02 0.0047329064 0.000000e+00
#> 4 1011 5.214638e-02 5.520133e-02 7.600797e-01 0.1325725492 0.000000e+00
#> 5 1008 6.707808e-01 1.917867e-01 1.374325e-01 0.0000000000 0.000000e+00
#> 6 2027 7.297028e-01 1.509212e-01 5.958378e-02 0.0284345396 3.135771e-02
#> 7 3010 7.689519e-01 1.789950e-01 1.700555e-02 0.0011966137 3.385091e-02
#> 8 1011 3.722963e-03 9.508964e-01 2.814040e-02 0.0020866235 1.515359e-02
#> 9 4002 1.000000e+00 0.000000e+00 0.000000e+00 0.0000000000 0.000000e+00
#> 10 2002 3.032605e-02 1.082159e-02 9.436752e-01 0.0124580620 2.719115e-03
#> 11 3000 5.338374e-01 3.903416e-01 6.427546e-03 0.0051871643 6.420623e-02
#> 12 3003 2.430521e-02 3.005071e-02 4.393773e-02 0.8884683982 1.323796e-02# Join bisg and shape file
bisg_sf <- dplyr::left_join(gwin_fulton_shape, bisg_dfsub, by = "BLOCKCE10")Plot Map of Proportion of Black Voters
# Plot choropleth map of race/ethnicity predictions for Fulton and Gwinnett counties
plot(bisg_sf["pred.bla"], main = "Proportion of Black Voters identified by BISG")
Plot Map of Proportion of White Voters
plot(bisg_sf["pred.whi"], main = "Proportion of White Voters identified by BISG")
Plot Map of Proportion of Hispanic Voters
plot(bisg_sf["pred.his"], main = "Proportion of Hispanic Voters identified by BISG")
Plot Map of Proportion of Asian Voters
plot(bisg_sf["pred.asi"], main = "Proportion of Asian Voters identified by BISG")