Skip to contents

Clean and standardize city and state fields

Source: R/augment.R
np_clean_geo.Rd

Normalizes city and state in place. Two engines:

Usage

np_clean_geo(
  data,
  engine = c("rules", "campfin"),
  aliases = np_city_aliases(),
  zip_ref = NULL,
  max_dist = 2
)

Arguments

data

A data frame with city / state (and zip5 for zip_ref).

engine

"rules" or "campfin".

aliases

Whole-name city alias map (see np_city_aliases()).

zip_ref

Optional zip5-keyed crosswalk for misspelling correction.

max_dist

Max edit distance to treat a city as a correctable near-miss.

Value

data with city / state cleaned in place.

Details

  • "rules" (default, no dependency) — upper-case and squish, expand the common abbreviated prefixes (ST->SAINT, FT->FORT, MT->MOUNT), apply whole-name city aliases (NYC->NEW YORK CITY), and coerce state to a two-letter code.

  • "campfin" — delegate city/state normalization to the campfin package (normal_city() / normal_state()), which validate against USPS and Census place lists (expanding abbreviations, stripping trailing state tokens, and nulling known-invalid entries). The alias map is still applied on top.

Correcting arbitrary misspellings (e.g. "PHOENex" -> "PHOENIX") needs a reference of valid place names. Supply zip_ref (a zip5-keyed crosswalk with a city column, e.g. from np_load_zipcode_db()): a missing city is filled from its ZIP, and a near-miss (edit distance <= max_dist) is snapped to the ZIP's canonical city.