---
title: "Workflow with tidycensus"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{2) Workflow with tidycensus}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---




``` r
library(zctaCrosswalk)
library(tidycensus)
library(dplyr)
```

`zctaCrosswalk` was designed to work well with the `tidycensus` package. 
`tidycensus` is currently the most popular way to access Census data in R.
Here is an example of using it to get Median Household Income on
all ZCTAs in the US:


``` r
zcta_income = get_acs(
  geography = "zcta",
  variables = "B19013_001",
  year      = 2021)
#> Getting data from the 2017-2021 5-year ACS

head(zcta_income)
#> # A tibble: 6 × 5
#>   GEOID NAME        variable   estimate   moe
#>   <chr> <chr>       <chr>         <dbl> <dbl>
#> 1 00601 ZCTA5 00601 B19013_001    15292  1299
#> 2 00602 ZCTA5 00602 B19013_001    18716  1340
#> 3 00603 ZCTA5 00603 B19013_001    16789   966
#> 4 00606 ZCTA5 00606 B19013_001    18835  2837
#> 5 00610 ZCTA5 00610 B19013_001    21239  1919
#> 6 00611 ZCTA5 00611 B19013_001    17143 10456
```

Note that `?get_acs` returns data for all ZCTAs in the US. It does 
not provide an option to get data on ZCTAs by State or County. And the dataframe it returns 
does not provide enough metadata to allow you to do this subselection yourself.

A primary motivation for creating the `zctaCrosswalk` package was to support this 
type of analysis. Note that `?get_acs` returns the ZCTA in a column called `GEOID`. We 
can combine this fact with `?dplyr::filter`, `?get_zctas_by_county` and 
`?get_zctas_by_state` to subset to any states or counties we choose.

Here we filter `zcta_income` to ZCTAs in San Francisco County, California:


``` r
nrow(zcta_income) 
#> [1] 33774

sf_zcta_income = zcta_income |>
  dplyr::filter(GEOID %in% get_zctas_by_county("06075"))
#> Using column county_fips

nrow(sf_zcta_income)
#> [1] 30
head(sf_zcta_income)
#> # A tibble: 6 × 5
#>   GEOID NAME        variable   estimate   moe
#>   <chr> <chr>       <chr>         <dbl> <dbl>
#> 1 94102 ZCTA5 94102 B19013_001    55888  8518
#> 2 94103 ZCTA5 94103 B19013_001    93143 19514
#> 3 94104 ZCTA5 94104 B19013_001    42591 34706
#> 4 94105 ZCTA5 94105 B19013_001   244662 44963
#> 5 94107 ZCTA5 94107 B19013_001   164289 16291
#> 6 94108 ZCTA5 94108 B19013_001    65392  9547
```

## Mapping the Result

A primary motivation in creating this workflow (and indeed, this package) was
to create demographic maps at the ZCTA level for selected states and counties.
If this interests you as well, I encourage you to copy the below code into R and view the output yourself. (Unfortunately, R package vignettes do not seem to handle map output from the `mapview` package well). This is a powerful and elegant pattern for visualizing ZCTA demographics in R:


``` r
library(zctaCrosswalk)
library(tidycensus)
library(dplyr)
library(mapview)

all_zctas = get_acs(
  geography = "zcta",
  variables = "B19013_001",
  year      = 2021,
  geometry  = TRUE)

filtered_zctas = filter(all_zctas, GEOID %in% get_zctas_by_county(6075))

mapview(filtered_zctas, zcol = "estimate")
```