US Population Data — pep-2025-r1
Snapshot assembled: September 30, 2026. Census source vintage: 2025.
Observation dates: July 1 of 2020–2025. Boundary vintage: January 1, 2025.
National coverage: 50 states and District of Columbia; excludes Puerto Rico and other territories.

Coverage: 1 nation, 50 states, 1 district, 3,144 counties/equivalents, 19,479 whole incorporated-place records.
Place source selection: SUMLEV 162 and functional status A or B, plus source-published F balance records. Balance profiles retained with labels; excluded from ordinary-place rankings.
County/place parts, consolidated-city totals and minor civil divisions are not independently ranked.
State-specific files include several overlapping geographic levels. Never sum those rows as if they formed a disjoint region.

LONG: one geography / July 1 date observation per row. Full source metadata included.
WIDE: one geography per row, six population columns. Year/date conventions are in the column names.
JSON: normalized geography records, annual values, source selection, land area, relationship codes, latest change components where supported.
Original raw source CSVs retained with exact downloaded bytes.
Census history: separate census-count table (1910–2020), historical definitions; not standardized boundaries.

Official codes must be imported as TEXT in spreadsheets to preserve leading zeroes.
Quoted CSV codes alone do not guarantee spreadsheet preservation. Internal IDs have type prefixes.
Units: people; land area square miles; density people per square mile.
Blank uncertainty is unavailable/not supplied, not zero. PEP files do not provide ACS-style margins of error.
Missing land joins suppress density; population zero is retained as a real zero.

Transformations (version 1): parse original CSV encoding as cp1252; keep leading-zero codes;
select geographic summary levels; derive state/place-part relationships; join 2025 Gazetteer on official codes;
convert square meters to square miles using 2,589,988.110336; do not round before calculations.
Absolute change = end-start; percentage = 100*(end/start-1), with positive start;
annualized = 100*((end/start)^(1/elapsed_years)-1), positive populations and interval;
density = population/compatible land area. Competition ranks share positions on ties.

Validation: duplicate codes/routes; nonnegative annual totals; state+DC national totals;
county totals for every state; births-deaths natural change; natural + domestic + international + residual reconciliation.
Source publication months: January 2026 (nation/state), March 2026 (county), May 2026 (place).
Retrieval and website snapshot date: September 30, 2026. See release.json and SHA256SUMS.txt.

Usage: Census public statistical data are federal works generally not subject to U.S. copyright
(Census policy DS027, https://www2.census.gov/foia/ds_policies/ds027.pdf).
Retain Census attribution and identify our transformations; original site text and branding have separate terms.
No competitor data or NHGIS collection is included. Independent website, not Census-endorsed.

Updates: a new vintage or substantive correction gets a new release ID. Never overwrite this snapshot.
Archived comparisons must pin the release ID and dates. Review current geographic coverage before combining releases.
Sources: https://uspopulationdata.com/sources/
Methods: https://uspopulationdata.com/methodology/
Contact/corrections: info@admk.com
