19-apis

Author

Affiliation

Prof Amanda Luby

Carleton College
Stat 220 - Spring 2025

TidyCensus

Install tidycensus if not on maize:

install.packages("tidycensus")

library(tidycensus)
library(tidyverse)

Example call:

acs_mn_2020 <- tidycensus::get_acs(
    year = 2020,
    state = "MN",
    geography = "county",
    variables = c("B01003_001", "B19013_001"),
    output = "wide",
    geometry = TRUE
)

Storing your API key

Create a new text file in the same folder as your .rmd
Copy and paste your census key into the empty file
Save the file as census_api_key.txt

my_key <- readLines("census_api_key.txt")

and tell tidycensus what your API key is with:

census_api_key(my_key)

Do not commit and push census_api_key.txt to github

acs_mn_2020 |>
  mutate(
    name = str_remove(NAME, ", Minnesota"),
    name = str_remove(name, " County")
  ) |> 
  ggplot(aes(x = B19013_001E, 
                          xmin = B19013_001E - B19013_001M, 
                          xmax = B19013_001E + B19013_001M, 
                          y = fct_reorder(name, B19013_001E))) + 
  geom_point() + 
  geom_errorbarh() + 
  labs(
    x = "",
    y = ""
  )

Links to more info about census

https://data.census.gov/

https://www.census.gov/acs/www/data/data-tables-and-tools/subject-tables/

Your turn

Search for two new variables using load_variables (Can you explain what they are?)
Run another call to tidycensus::get_acs using your two variables

httr2

Install if not already installed:

install.packages("httr2")

library(httr2)

Example call

hmong_state_request <- request("https://api.census.gov/data") %>% 
    req_url_path_append("2019") %>% 
    req_url_path_append("acs") %>% 
    req_url_path_append("acs1") %>% 
    req_url_query(get = c("NAME", "B02015_009E", "B02015_009M"), `for` = I("state:*"), key = census_api_key, .multi = "comma")

hmong_state_response <- req_perform(hmong_state_request)

hmong_state_tbl <- hmong_state_response %>%
  resp_body_json(simplifyVector = TRUE) %>%
  janitor::row_to_names(1) %>%
  as_tibble()

hmong_state_tbl

Your Turn

Edit the httr code to access a new variable of your choice
Make a httr request to access the 1-year ACS data from 2018, 2019, 2021, and 2022. Make sure to save your results from each call!
Combine all years into a single dataset
Make a time series plot with your chosen variable on the y-axis, year on the x-axis, colored by state.
- You may want to first filter to only a few states
- You will need to do some cleaning of the data

--- title: "19-apis" format: html --- # TidyCensus Install `tidycensus` if not on maize: ```{r} #| eval: false install.packages("tidycensus") ``` ```{r} library(tidycensus) library(tidyverse) ``` ## Example call: ```{r} #| eval: false acs_mn_2020 <- tidycensus::get_acs( year = 2020, state = "MN", geography = "county", variables = c("B01003_001", "B19013_001"), output = "wide", geometry = TRUE ) ``` ## Storing your API key 1. Create a new text file in the same folder as your .rmd 2. Copy and paste your census key into the empty file 3. Save the file as `census_api_key.txt` ```{r} #| eval: false my_key <- readLines("census_api_key.txt") ``` and tell `tidycensus` what your API key is with: ```{r} #| eval: false census_api_key(my_key) ``` *Do not commit and push `census_api_key.txt` to github* ```{r} #| eval: false acs_mn_2020 |> mutate( name = str_remove(NAME, ", Minnesota"), name = str_remove(name, " County") ) |> ggplot(aes(x = B19013_001E, xmin = B19013_001E - B19013_001M, xmax = B19013_001E + B19013_001M, y = fct_reorder(name, B19013_001E))) + geom_point() + geom_errorbarh() + labs( x = "", y = "" ) ``` ## Links to more info about census <https://data.census.gov/> <https://www.census.gov/acs/www/data/data-tables-and-tools/subject-tables/> ## Your turn - Search for two new variables using `load_variables` (Can you explain what they are?) - Run another call to `tidycensus::get_acs` using your two variables # httr2 Install if not already installed: ```{r} #| eval: false install.packages("httr2") ``` ```{r} library(httr2) ``` ## Example call ```{r} #| eval: false hmong_state_request <- request("https://api.census.gov/data") %>% req_url_path_append("2019") %>% req_url_path_append("acs") %>% req_url_path_append("acs1") %>% req_url_query(get = c("NAME", "B02015_009E", "B02015_009M"), `for` = I("state:*"), key = census_api_key, .multi = "comma") ``` ```{r} #| eval: false hmong_state_response <- req_perform(hmong_state_request) ``` ```{r} #| eval: false hmong_state_tbl <- hmong_state_response %>% resp_body_json(simplifyVector = TRUE) %>% janitor::row_to_names(1) %>% as_tibble() hmong_state_tbl ``` ## Your Turn - Edit the `httr` code to access a new variable of your choice - Make a `httr` request to access the 1-year ACS data from 2018, 2019, 2021, and 2022. Make sure to save your results from each call! - Combine all years into a single dataset - Make a time series plot with your chosen variable on the `y-axis`, year on the `x-axis`, colored by `state.` - You may want to first filter to only a few states - You will need to do some cleaning of the data