arbonet

Ingests CDC ArboNET historic county-level arboviral disease surveillance data for PopHIVE.

Six domestic arboviruses, each from its CDC historic-data dashboard:

Code	Virus	Years
wnv	West Nile virus	1999–2025
eee	Eastern equine encephalitis virus	2003–2025
jcv	Jamestown Canyon virus	2011–2025
lac	La Crosse virus	2003–2025
sle	St. Louis encephalitis virus	2003–2025
pow	Powassan virus	2004–2025

Output

Both files are wide, keyed on geography + time (YYYY-12-31, annual).

• standard/data_county.csv.gz — 5-digit county FIPS, plus a national ("00") row per year so county series can be charted against the national trend. Filter on nchar(geography) before summing across rows.
• standard/data_state.csv.gz — 2-digit state FIPS plus national ("00").

Measure columns are arbonet_<code>_<measure>:

Measure	Meaning
_cases	Total reported human disease cases (neuroinvasive + non-neuroinvasive)
_neuro_cases	Neuroinvasive cases (meningitis, encephalitis, acute flaccid paralysis)
_nonhuman_activity	1 if veterinary/mosquito/bird/sentinel activity was reported, else 0. Not produced for pow, whose dashboard reports human cases only.
arbonet_wnv_blood_donors	Presumptive viremic blood donors found by blood-supply screening. WNV only.

See measure_info.json for full definitions and caveats.

Usage

source("ingest.R")

No API key required, but note: the dashboard pages are JavaScript visualizations with no download endpoint. ingest.R reads the flat CSVs that back them under www.cdc.gov/wcms/vizdata/, resolved from each page's cdc-open-viz config JSON. www.cdc.gov also sits behind Akamai and answers most non-browser clients with a 403 — download.file(method = "libcurl") gets through where httr and plain curl do not.

Interpreting the data

• Cases are reported by county of residence, not where exposure occurred.
• Non-neuroinvasive disease is substantially under-reported. CDC advises against using total case counts to compare activity between locations or over time; use _neuro_cases for that.
• Non-human surveillance is not standardized. A 0 means nothing was reported to CDC, not that there is no risk.
• The most recent year is preliminary and subject to revision.
• Connecticut changes geography mid-series — counties through 2022, planning regions from 2023 — so its FIPS codes shift.
• Unallocated records (state FIPS + 999, used when county of residence is unknown) are held out of the county rows but retained in the state and national totals, so those still reconcile with CDC's published figures. The ingest logs which codes it held out.
• Counts are passed through unsuppressed, exactly as CDC publishes them. County values of 1–2 cases are common; no further suppression or imputation is applied here.

Consumed by

Intended to be pulled by PopHIVE/Ingest directly from GitHub, following the same pattern as PopHIVE/hud-chas, PopHIVE/bureau-labor-statistics, and PopHIVE/usda-food-access.