census_geocoder_multi_batch: Function to geocode a large (data frame) of addresses in...

In sakrejda/workflow: A sharable but idiosyncratic workflow

Function to geocode a large (data frame) of addresses in street/city/state/zip format

Description

The data here must still be able to fit in memory easily, data duplication does occur.

Usage

census_geocoder_multi_batch(
  data,
  street = street,
  city = city,
  state = state,
  zip = zip,
  endpoint = census_geocoder_api_endpoint(),
  returntype = "locations",
  benchmark = purrr::flatten(purrr::keep(census_geocoder_api_benchmarks(),
    ~.x$isDefault)),
  vintage = purrr::flatten(purrr::keep(census_geocoder_api_vintage()$vintages,
    ~.x$isDefault)),
  cache_dir = workflow::build_dir("census-geocoder-cache"),
  batch_size = 100,
  lib_paths = .libPaths()
)

Arguments

street	bare symbol for street column
city	bare symbol for city column
state	bare symbol for state column
zip	bare symbol for zip code column
cache_dir	optional path to file to use to cache results to avoid submitting them multiple times. Can be shared between calls or it will be based on a hash of the batch input data.
batch_size	how many records should be sent to a worker at one time
batch	data frame containing columns with parts of the address
...	other arguments to the batch geocoder

Value

list of geocoded components and resulting tables with some simplified output

sakrejda/workflow documentation built on Nov. 18, 2024, 3:02 a.m.