County data desk
How does this desk build a county number? From two public files, one column at a time.
Every figure published by Senior Care Map STL traces to a named file, a named column and a pull date. This page lists all of them, together with the things this desk deliberately does not publish and why.
The two inputs in one box
- Population
- U.S. Census Bureau, Population Estimates Program, Vintage 2024 county characteristics by age and sex. Columns POPESTIMATE, AGE65PLUS_TOT and AGE85PLUS_TOT. Rows coded YEAR 6.
- Businesses
- U.S. Bureau of Labor Statistics, Quarterly Census of Employment and Wages, 2024 Q4 county files. Columns qtrly_estabs and month3_emplvl, private ownership, NAICS 621610.
- Pull date
- September 10, 2026. Both were read directly from the agencies' own hosts on that date, not from a mirror or an aggregator.
- Licence
- Public domain. Both are works of the United States government.
- Computed here
- Two figures on this site are computed rather than published: the share of a county's population aged 65 and over, and home care establishments per 1,000 residents aged 65 and over. Both formulas are stated on the pages that use them.
The population file
The Census Bureau publishes one county characteristics file per state in each estimates vintage. This desk reads two of them: cc-est2024-agesex-29.csv for Missouri and cc-est2024-agesex-17.csv for Illinois. The header of each was verified live on the pull date rather than assumed from a prior run.
Three columns are used. POPESTIMATE is total resident population. AGE65PLUS_TOT is residents aged 65 and over. AGE85PLUS_TOT is residents aged 85 and over, and it is a subset of the 65 and over column, not an addition to it. Only rows where SUMLEV is 050, the county summary level, are read.
These are estimates. The Bureau builds them from the 2020 census base and updates them with administrative records covering births, deaths and migration. They are not a fresh count of anybody, and the smallest counties carry the widest relative uncertainty. The agency's own estimates methodology documentation is the authority on how they are produced.
What YEAR 6 means
Each county appears in the file several times, once per reference period in the vintage series, distinguished by a YEAR code. In the 2020 to 2024 vintage the codes run 1 through 6. Code 1 is the April 1, 2020 census base. Codes 2 through 6 are the July 1 estimates for 2020, 2021, 2022, 2023 and 2024 in order.
This desk reads YEAR 6 and only YEAR 6, so every population figure on this site describes July 1, 2024. Mixing YEAR codes inside one table is the most common way to produce a county comparison that looks coherent and is not, so the filter is applied once at load and never relaxed.
A pull date of September 10, 2026 and a reference date of July 1, 2024 are two different facts, and both are printed on every page. The lag is the Census Bureau's publication cycle, not a stale file.
The county list
Regional pages cover the 15 counties of the St. Louis MO-IL metropolitan statistical area as delineated by the Office of Management and Budget. In Missouri: Franklin, Jefferson, Lincoln, St. Charles, St. Louis and Warren counties, plus the independent city of St. Louis, which is a county equivalent and appears as its own row. In Illinois: Bond, Calhoun, Clinton, Jersey, Macoupin, Madison, Monroe and St. Clair counties.
The statewide page covers all 115 Missouri county equivalents, meaning the 114 counties plus the city of St. Louis. No county is filtered out on size, and no county is added from outside those two definitions.
The independent city is the one row that trips most county comparisons. It is not part of St. Louis County, its figures are not included in the St. Louis County row, and adding the two together would double count nobody but would merge two jurisdictions that make separate decisions. They are kept separate everywhere on this site.
The business file
Establishment and employment counts come from the Bureau of Labor Statistics Quarterly Census of Employment and Wages, 2024 Q4, read one county file at a time from the agency's county data interface. The rows read are private ownership, aggregation level 78, industry code 621610, home health care services. The columns are qtrly_estabs for establishments and month3_emplvl for employment in the third month of the quarter.
NAICS 621610 mixes medical and non-medical home care into a single industry code. A Medicare certified skilled nursing agency and a company that sends a caregiver to help with bathing, meals and errands are both classified there. No public column separates them, so no page on this site claims to count non-medical providers specifically. The ratio measures registered business presence in an industry code, which is a narrower claim than it might first appear.
Two absences in that file are different from each other and are treated differently here. Disclosure code N means BLS is withholding a figure to protect an identifiable employer; where that applies, the establishment count is still published and only employment is withheld, and this site shows employment as withheld. A county with no 621610 row at all is a different case: since suppressed rows still carry an establishment count, an entirely missing row reads as nothing reported rather than as something hidden. Those counties are shown as no row published and no ratio is computed for them.
What is deliberately not published here
Municipality figures. This desk publishes county totals and nothing smaller. Place level population by age comes from the American Community Survey rather than the population estimates programme, and that survey's data interface now requires a key. Requested without one it answers HTTP 200 and returns an HTML error page, which a naive parser accepts as a successful response and turns into numbers that look real. There is no safe way to publish town figures on that footing, so Ballwin, Chesterfield, Valley Park and every other municipality are absent from this site by decision, not by oversight.
Anything about health. Population counts describe who lives where by age. They do not identify need, diagnosis, disability, independence or clinical status, and no page here converts a count into a care requirement. Nothing on this site is medical, diagnostic or treatment information of any kind.
Provider quality, licensure or capacity. Neither source file carries any of it. This desk does not rate, score, certify or vouch for any provider, does not operate a referral service, and does not match a family to a company.
Projections. Nothing here is forecast forward. Every figure is the published value for its stated reference date.
Editorial rules this desk holds to
A page is published only when the underlying file supports it. A ranking needs at least six rows carrying real values and a table needs at least ten, and a page that cannot clear its own bar is not published in a thinner form.
Figures are reproduced as the source file gives them. Percentages computed here are rounded for display and the rounding is visible in the column heading. Where a computed figure is this desk's own construction rather than an agency's, the page says so in the methodology box.
One placement on this site is paid. New Plan Care is a client of the publisher, OwnersFirm, which is paid for that slot. It appears in a bordered, labelled box that sits outside every ranked table, it holds no rank position, and it appears in no ItemList structured data. No county on any page was ordered, filtered or worded in its favour, and removing the placement would change no number on this site.
Corrections are made in place, and the revision date at the top of this page moves when they are. If a figure here disagrees with the source file, the source file is right and this desk wants to know.