The core source is the Clark County Clerk's marriage statistics page, which links annual statistical CSV files. The Clerk asks users to reference that page and says files may be updated to reflect document corrections. We keep acquisition dates and hashes internally and publish only aggregates. This site is independent of the Clerk and the other source publishers.
Dates, totals and the incomplete file
The date is the license issue date. It is never used as a ceremony date. Each dated row contributes a license to the relevant mailing-address group. Rows without a license date are dropped, including padding in some files. The files use mixed text encodings, recorded internally for each acquisition. Strict decoding is used, with no character silently replaced during counting.
The 2022 file stops on 22 December and contains 77,006 dated rows. The Clerk's dashboard gives 79,279 licenses for that year. Every 2022 breakdown here is labeled as coverage to 22 December. We do not scale its origin, age or month distributions up to the dashboard total.
Annual county totals come from the dashboard, which separates licenses from marriages. The fetched response exposes 2022 through 2026. For 2007 through 2021, annual license figures are labeled 'licenses in the Clerk's data file'. Each file was checked to start on January first and end on December thirty-first with all twelve months present. Earlier dashboard marriage totals were not obtained. State and country breakdowns come from the statistical files. A live dashboard may change after corrections or later certificate filings.
Addresses and entities
The mailing state, country and locality classify the license. They do not identify nationality, immigration status or the usual residence of each party. Case, accents, punctuation and repeated whitespace are normalized. Country aliases are resolved against the UN M49 list of countries and areas. Unresolved labels remain unclassified. A US territory in the state field retains its territory label.
England, Scotland, Wales and Northern Ireland combine into the UK entity, while keeping their nation labels in the breakdown. Entries labeled only United Kingdom stay unspecified within that entity. Ireland is separate. Crown Dependencies are not added to the UK. Locality matches use one exact normalized Census place name within a confirmed state; ambiguous and unmatched labels receive no city page. Postal place names do not establish municipal residence.
Age, history and gender
Age is license year minus birth year, grouped into five-year bands. It is the age reached that year, not exact age at license issuance. Missing, nonnumeric or implausible values remain unclassified. The quality screen accepts derived ages from fifteen through one hundred. The distribution includes both parties and all marriages. The ACS age estimates cover first marriages.
Both-first-marriage licenses require both reported marriage numbers to indicate a first marriage. At-least-one-repeat licenses require a reported later marriage number for either party. Unknown histories stay unclassified. These categories describe licenses, not previous spouses or divorce outcomes.
Gender fields are largely missing before 2015 and mix abbreviated and full labels in the transition. Abbreviated male and female labels are normalized for the quality check. Later files pass a completeness screen for the binary fields, but that does not verify historical accuracy or identity. The recorded-gender page publishes the eligible annual distributions, with other or missing values unclassified. Its completeness threshold is 99%. The recorded binary pairings supply no sexual-orientation classification or complete LGBTQ marriage total.
Statistical joins
State first-marriage ages come from Census ACS B12007, 2024 single-year estimates, with margins of error displayed. State marriage counts are estimates from the CDC 2024 provisional occurrence rate multiplied by Census population. A state occurrence denominator excludes weddings its residents hold elsewhere and can include visitors. It is not a count of resident couples choosing between Nevada and home.
Adult rates use Census resident adults aged eighteen and over in 2025. Distance uses the Census Gazetteer, first matching postal-area centroids and then unique place centroids. The endpoint is the Las Vegas place centroid. The median covers matched records only. Countywide distance coverage uses licenses with a recognized US-state or DC address as its denominator, excluding US labels with an unresolved state. These are great-circle miles, not road distances or evidence of transport used. No postal codes are exported.
Gretna figures come from NRS table MCPT.05. The published comparison uses marriages where neither party resides in Scotland, divided by all marriages at Gretna. It does not reuse the workbook's Scotland-wide percentage column as a Gretna percentage. The Gretna story uses this residence classification alongside UK mailing-address licenses.
Heat and the licensing calendar
NOAA GHCN-Daily station USW00023169, named MCCARRAN INTL AP in the official inventory, supplies daily maximum temperature. Missing and quality-flagged temperatures are excluded. The heat join contains 6,931 covered issue dates. Missing final days are excluded from the calculation. The heat calculation subtracts the average temperature and license total within each year, month and weekday group before computing their correlation.
Calendar-day and weekday averages include zero-license days inside each file's checked date span. Missing late-December days and leap-day coverage retain their actual contributing years. The date picker uses local aggregate data and has the same values as its no-script fallback table.
Privacy before publication
An entity needs at least 100 licenses in the recent window for a page, US states need 500, and city candidates need 250. A nonzero displayed count below 10 is replaced with 'fewer than 10'. Complementary cells are withheld too. Exact hidden values are removed before JSON is written, and their rates are not released. The same suppressed partition feeds charts, tables and downloads.
No record identifiers, individual names, postal-code cells or row-level downloads are published. Locality is never crossed with an exact day, age and marital history. State city tables use only the recent-window count. City age, month and history blocks are omitted in full when any nonzero cell is below the floor; population rates are omitted without a verified Census place match. Explicit zeros are published. The privacy gate checks the output files for small cells and hidden values.
Downloads and credits
Download the current release aggregates, the historical calendar or the Gretna aggregates. The JSON uses null for a suppressed cell, with its reason. These are this site's calculations under CC BY 4.0; source publishers retain their own rights. The NRS data is Crown copyright, reused under the Open Government Licence v3.0. CDC data is available free from the agency, and its use here implies no endorsement by CDC, HHS or the US government.
- Census ACS median age at first marriage summary file
- Census state Gazetteer
- Census place Gazetteer
- Census ZCTA Gazetteer (centroids used internally only)
- Census adult population estimate
- Census resident population estimates
- CDC/NCHS marriage rates by state
- UN countries and areas
- National Records of Scotland, marriage time series
- Saved CDC state-rate workbook
- Census incorporated-place population estimates
- NOAA airport daily observations