HomeCountiesArticlesMethodologyContactField definitions for the national CSV (caeri-national.csv) and the per-county JSON files.
| Column | Meaning |
|---|---|
fips | 5-digit county FIPS code (stable identifier). |
county, state_usps | County name and 2-letter state. |
national_caeri | CAERI composite score (0–100), national basis. |
national_percentile, national_rank | Standing among all scored U.S. counties. |
national_tier | Relative-standing band by national percentile: Top 10% / 75th–90th / 50th–75th / 25th–50th / Bottom 25%. A ranked position among all U.S. counties, not a condition assessment. |
national_confidence | High / Medium / Low — trust in the estimate (source agreement, disclosure imputation, county size). Distinct from the tier band above. |
p1–p5 | Pillar national percentiles: P1 direct occupational exposure, P2 economic concentration, P3 adaptive capacity (inverted), P4 regional buffer (inverted), P5 fiscal sensitivity. |
instate_caeri, instate_percentile, instate_rank, instate_confidence | The same score, percentile, rank, and confidence computed within the county's own state. There is no in-state tier band (removed v0.3.1): percentile bands are incoherent in small states (DC has 1 county, DE 3). Use the rank — it is strictly more informative. |
Every county carries both a national standing (vs. all U.S. counties) and an in-state standing (vs. its own state's counties). They can differ sharply; both are shown so a reader can pick the comparison that fits. Only the national basis carries a tier band; the in-state basis is reported as rank and percentile.
Tier bands are relative-standing cuts on the national percentile, not condition assessments: Top 10% (90th percentile and above), 75th–90th, 50th–75th, 25th–50th, and Bottom 25%. They apply to the national basis only. P3 and P4 are resilience pillars (higher = more resilient) and are inverted inside the composite. Pillar weights and the full construction are in the methodology.
Each county page links a JSON file with a metadata block (identifier,
canonical URL, methodology version, sources, license), a scores block
(national and in-state bases + pillar percentiles), and any
data_quality_flags.