RcensusPkg – deandevl

Introduction

The goal of RcensusPkg is to provide easy access to the US Census Bureau’s datasets and collection of TIGER/Line Shapefiles providing plot geometries for states, counties, roads, landmarks, water, enumeration tracts/blocks for the entire United States. The only requirement is for the user to apply for and obtain a free access key issued from the Bureau. See Guidance for Developers for additional information.

The example below illustrates a simple workflow for downloading a dataset, merging the data with shapefile geometries, and plotting the merge to create a choropleth map.

Installation

You can install the development version of RcensusPkg from GitHub with:

# install.packages("pak")
pak::pak("deandevl/RcensusPkg")

Using devtools::install_github():

devtools::install_github("deandevl/RcensusPkg")

Also for this example we will need devtools::install_github("deandevl/RplotterPkg")

Example

US Census Bureau API key

All Census Bureau API requests require an access key. Sign-up for a key is free and can be obtained here. The functions will check for a global setting of the key via Sys.getenv(CENSUS_KEY). Run usethis::edit_r_environ() and edit your .Renviron file with the line: CENSUS_KEY=your key to create the global association.

Setup

We will be using the following packages.

library(httr)
library(jsonlite)
library(stringr)
library(data.table)
library(withr)
library(sf)
library(kableExtra)
library(ggplot2)
library(RplotterPkg)
library(RcensusPkg)

A look at the US Census Bureau’s community resilience estimates (CRE) database.

Among the list of available API, there is the Community Resilience Estimates based on such factors as:

Income-to-Poverty Ratio (IPR) < 130 percent
Single or zero caregiver household
Aged 65 years or older
No health insurance coverage
No vehicle access (Household)
Disability, at least one serious constraint to significant life activity
No one in the household is employed full-time, year-round
Households without broadband internet access
Unit-level crowding with >= 0.75 persons per room
No one in the household has received a high school diploma
No one in the household speaks English very well

The factors are used to estimate the number of people with:

0 risk factors (Low risk)
1-2 risk factors (Medium risk)
3 or more risk factors (High risk)

The workflow for using RcensusPkg is the following:

Get a database name recognized by the API. Use RcensusPkg::get_dataset_names() filtering for a resilience title and vintage of 2022.

datasets_dt <- RcensusPkg::get_dataset_names(
  vintage = 2022,
  filter_title_str = "resilience"
)

Table 1: resilience datasets
name	vintage	title
cre	2022	Community Resilience Estimates
crepuertorico	2022	Community Resilience Estimates for Puerto Rico

The returned dataframe shows a dataset name for CRE 2022 as (surprise) cre.

Get the variable names available for the cre dataset.

cre_var_names_dt <- RcensusPkg::get_variable_names(
  dataset = "cre",
  vintage = 2022
) |> 
  _[, .(name, label)]

Table 2: Variable names from the CRE dataset
name	label
COUNTY	Geography
GEOCOMP	GEO_ID Component
GEO_ID	Geographic Identifier
NATION	Geography
POPUNI	Population Universe
PRED0_E	Estimated number of individuals with zero components of social vulnerability
PRED0_PE	Rate of individuals with zero components of social vulnerability
PRED12_E	Estimated number of individuals with one-two components of social vulnerability
PRED12_PE	Rate of individuals with one-two components of social vulnerability
PRED3_E	Estimated number of individuals with three or more components of social vulnerability
PRED3_PE	Rate of individuals with three or more components of social vulnerability
STATE	Geography
SUMLEVEL	Summary Level code
TRACT	Geography
for	Census API FIPS for clause
in	Census API FIPS in clause
ucgid	Uniform Census Geography Identifier clause

We are interested in the percentage of individuals with three or more vulnerabilities (PRED3_PE)

Get the regions available for the CRE dataset.

cre_regions_dt <- RcensusPkg::get_geography(
  dataset = "cre",
  vintage = 2022
)

Table 3: Regions from the CRE dataset
name	geoLevelDisplay
us	010
state	040
county	050
tract	140

So we can get CRE estimates from the entire US, state, county, and tract enumeration levels. We are interested in the counties for the state of Florida.

Download the data:

florida_fips <-  usmap::fips(("FL"))

florida_cre_dt <- RcensusPkg::get_vintage_data(
  dataset = "cre",
  vintage = 2022,
  vars = "PRED3_PE",
  region = "county",
  regionin = paste0("state:", florida_fips)
) |> 
  _[, PRED3_PE := as.numeric(PRED3_PE)] |> 
  data.table::setnames(old = "PRED3_PE", new = "CRE_GE_3")

Table 4: Florida county percent risk from the CRE dataset
NAME	CRE_GE_3	state	county	GEOID
Alachua County, Florida	19.29	12	001	12001
Baker County, Florida	22.93	12	003	12003
Bay County, Florida	20.05	12	005	12005
Bradford County, Florida	24.77	12	007	12007
Brevard County, Florida	20.67	12	009	12009
Broward County, Florida	22.44	12	011	12011
Calhoun County, Florida	31.52	12	013	12013
Charlotte County, Florida	27.53	12	015	12015

Merge the CRE county data with the Florida county shapefile geographies from the Bureau. Note that we need to establish a folder for receiving the shapefiles.

output_dir <- withr::local_tempdir()
if(!dir.exists(output_dir)){
  dir.create(output_dir)
}
express <- parse(text = paste0("STATEFP == ", '"', florida_fips, '"'))
cre_florida_sf <- RcensusPkg::tiger_counties_sf(
  output_dir = output_dir,
  vintage = 2022,
  general = TRUE,
  express = express,
  datafile = florida_cre_dt,
  datafile_key = "county"
)

Plot the choropleth map of the percent of individuals with 3 or more risk factors.

RplotterPkg::create_sf_plot(
  sf = cre_florida_sf,
  aes_fill = "CRE_GE_3",
  hide_x_tics = TRUE,
  hide_y_tics = TRUE,
  panel_color = "white",
  panel_border_color = "white"
)

Figure 1: Percent 3 or more risk factors among Florida counties

A more detailed example of the RcensusPkg workflow is available here.