6 Geocoding and Location Data
Learn how geocoding converts address descriptions into coordinates, how to prepare and review address matches, and how to manage uncertainty in location data.
and coordinates
converts a place description—usually a postal address—into coordinates such as latitude and longitude. uses coordinates to find a nearby human-readable address. Geocoded records can be mapped, compared with other spatial data, or assigned to administrative areas such as census tracts.
A result is an estimate based on the geocoder’s and matching method. It is not automatically proof that the address is valid or that the coordinate marks a building’s exact location.
Preparing address records
Well-prepared input improves the chance of a useful match. Keep a unique ID for each record so results can be joined back to the original data. When possible, separate address components such as house number, street name and type, unit, city, state or province, postal code, and country.
Use consistent formats and remove accidental extra spaces while preserving meaningful details, including directional prefixes such as N or South and unit numbers. Include enough locality information to distinguish similar street names or addresses; a country or state can be especially important when names repeat.
Keep both the original address and any cleaned or standardized version so that corrections remain traceable. Handle missing components explicitly rather than shifting columns in a batch file, and follow the selected service’s input format and regional conventions. For example, 4600 Silver Hill Rd, Washington, DC 20233 is easier to distinguish than 4600 Silver Hill. Postal-address rules and available address data vary by country, and adding a unit number does not guarantee that the geocoder can locate that unit precisely.
Matching and reviewing results
A geocoder compares an input with its address database or road network, resolves the best-supported location, and returns coordinates and other information. Depending on the service and input, a response may include a , address components, a result type, match indicators, and a or location-quality category. Some services return multiple candidate matches, others return one candidate, and some return no match. Match scores and labels are service-specific and should not be treated as directly comparable across providers.
Review each result rather than accepting every coordinate automatically:
Check whether a result was found. A no-match result means the service could not locate the address in its ; it does not by itself prove that the address does not exist.
Compare the returned address with the input. Check the house number, street, locality, region, and postal code for substitutions or omissions.
Check the result type and . A street-address result differs from a street, locality, or postal-code result. Some systems distinguish rooftop or address-level locations from interpolated or approximate ones.
Review partial or ambiguous matches. A partial match means only part of the request was resolved. It may point to a misspelling, an incomplete address, or more than one possible place. Use better input or manual review to resolve uncertainty.
Retain match information. Store the original input, , coordinates, result type, provider, and relevant dataset or benchmark version to support auditing and repeatability.
For example, if 123 Example Street returns a location for the street but not a specific house number, do not treat the point as a confirmed address-level location. A match to the right city but the wrong street is not a successful match for most address-level analyses.
Understanding accuracy limitations
Several limitations can affect how closely a returned coordinate represents the intended place:
: If a precise address point is unavailable, a geocoder may estimate a position along a road segment between known address locations. The coordinate can be near the address without marking its building or entrance.
Approximate or generalized results: A result may identify a street, neighborhood, postal area, or city rather than a particular property. The number of decimal places in a coordinate does not establish its real-world accuracy.
Incomplete, misspelled, or ambiguous input: Missing locality information, typographical errors, alternate street names, and duplicate street names can lead to partial, incorrect, or no matches.
Reference-data gaps and change: New developments, rural addresses, informal settlements, renamed streets, and recently changed postal information may be missing from a geocoder’s data or out of date.
Address versus location of interest: A postal address may identify a parcel, building, entrance, or mailing point. The appropriate choice depends on the task; a delivery study may need an entrance, while a neighborhood analysis may tolerate an approximate address point.
Coordinate and geographic-area uncertainty: A small positional error can put a point on the wrong side of a street or administrative boundary. Assigning census or other area data to a geocoded point inherits both coordinate error and limitations in the boundary data.
Provider and version differences: Different services, reference datasets, settings, and update dates can return different matches for the same input. Record the service and data version used, and recheck results when the work must be reproduced.
Quality control and responsible use
Map a sample of results over roads, imagery, or authoritative local address data when available. Flag no matches, partial matches, unexpected locality changes, and low- result types for review.
For high-stakes uses—such as emergency response, eligibility decisions, or property-level analysis—do not rely on a geocoded coordinate alone. Verify the address and location against an appropriate authoritative source. The level of review should reflect the consequences of a location error.