Funding & Governance¶
Every dataset this repo collects is published by the CDC, but almost none of it is generated there. Notifiable-disease counts, outbreak reports, enteric-pathogen isolates, behavioral-risk estimates, vaccination coverage, vital records — these are produced by state and local health departments and reported upward. What makes that reporting happen is money: CDC moves appropriations down to the states, mostly through cooperative agreements, and the surveillance those agreements fund is what surfaces, years later, as the tables the other pages here query.
This page traces that loop — federal dollars down, data back up — and enumerates the institutions in the middle. It is a governance map, not a budget: it names the funding mechanisms, not dollar figures.
The money flow¶
Pick a state to expand its real department and its county/local health departments,
pulled from the scraped directories below. Solid green edges are federal dollars flowing
down; dashed teal edges are the surveillance those programs generate, flowing back up as
the datasets this repo collects. The graph is a Svelte Flow
component compiled into the site from frontend/.
The three tiers¶
| Tier | Who | This repo's view |
|---|---|---|
| Federal | CDC (plus NCHS, ATSDR, and NCI for cancer) | The APIs and bulk files every other page queries |
| State | 50 state health departments + DC | Enumerated from the FSIS directory (state_departments.csv) |
| County / local | ~3,400 local health departments (LHDs) | Enumerated from the NACCHO directory (local_departments.csv) |
Federal surveillance mostly stops at the state line (see State & Local Sources), but the institutions that feed it run all the way down to the county. The two directories scraped here are the roster of that middle-and-bottom tier.
Mechanism → dataset¶
Each cooperative agreement or program funds a slice of state/local capacity, and that capacity is what produces a specific dataset. The correspondence is approximate — real appropriations are braided — but it is the honest shape of why each table exists.
| CDC mechanism | Tier funded | Surfaces in this repo as |
|---|---|---|
| ELC — Epidemiology & Laboratory Capacity | State epidemiology + public-health labs | NNDSS, NORS, BEAM, Lyme |
| PHEP — Public Health Emergency Preparedness | State + local ED/syndromic infrastructure | NSSP |
| BRFSS cooperative agreement | State behavioral-risk surveys | PLACES small-area estimates |
| NPCR — National Program of Cancer Registries | State central cancer registries | Cancer statistics (see note) |
| §317 immunization grants | State immunization programs | NIS coverage |
| EPHT tracking grants | State environmental-tracking programs | EPHT |
| NVSS vital-statistics cooperative | State vital-records offices | WONDER, WISQARS |
SEER is not CDC-funded
The cancer node is a deliberate simplification. NPCR (CDC) funds registries in most states and territories and feeds U.S. Cancer Statistics; SEER is a parallel program run by the NCI, covering a smaller set of registries. This repo collects SEER, so the graph points NPCR at the SEER node as the nearest cancer surface — the funder underneath is NCI, not CDC.
The scraped directories¶
Two authoritative, machine-readable rosters are refreshed on a schedule and committed to
data/processed/health_depts/.
State departments — FSIS¶
The USDA FSIS directory of State Departments of Public Health and Agriculture is one static table: for each of the 50 states + DC it lists the state's home portal, its public-health department, and its agriculture department, each with a link.
Local departments — NACCHO¶
The NACCHO Local Health Department Directory
lists ~3,400 county, city, and district health departments. The scraper walks all 50 states
+ DC (?lhd-state=XX) and captures name, address, phone, website, and social links for each.
uv run python -m health_depts local -f csv | head
uv run python -m health_depts all --out # write both directories + summary.json
Both directories together are ~2 MB — small enough to commit whole, in keeping with the
ingestion discipline that governs data/. They change slowly,
so the update_health_depts workflow refreshes
them monthly; the graph above rebuilds from summary.json on the next docs deploy (the
Svelte project in frontend/ bakes the data in at build time).
Caveats¶
- Mechanisms, not amounts. No dollar figures appear here. A data-backed version — real award amounts per state from USAspending / TAGGS — is a deliberate future extension, not part of this map.
- County labels are best-effort. The graph's county names are derived from department
names (
"Autauga County Health Department"→Autauga); non-standard names (city districts, combined health-and-human-services agencies) keep more of their full title. - Directories drift. FSIS links and NACCHO membership go stale; treat the scrape date shown in the graph footer as the as-of.