Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,152 @@
+# FDA Facility Compliance Intelligence
+**Version:** 1.0.0 | **Records:** 132,080 | **Price:** $3,500 | **Source:** FDA (public domain)
+
+## Dataset Summary
+**The dataset's core signal — whether a facility's inspections escalated to a Warning Letter — tracks FDA's own severity classifications: facilities whose most recent inspection was OAI have escalated to a Warning Letter 65.7% of the time, versus 2.5% for NAI — a ~26× relationship you can reproduce directly from this file (`GROUP BY most_recent_classification_code`, averaging `ever_escalated_to_wl`).**
+
+Complete facility-level regulatory intelligence for every FDA-regulated establishment with an inspection on record — 132,080 facilities. Each row rolls up a facility's entire FDA history: inspection outcomes (OAI/VAI/NAI), Warning Letters, recalls, import refusals, citation analytics, 510(k) device clearances, and a computed escalation summary, condensed into a transparent 0–100 risk score. Built for pharma/medtech regulatory affairs and quality (RA/QA) teams, supplier-qualification and auditing groups, due-diligence analysts, and investors who need to know a facility's compliance posture at a glance.
+
+Current through 2026 — the dataset includes FDA inspections from the most recent regulatory cycles
+(5,713 facilities inspected in 2026, 14,845 in 2025), not a historical snapshot. Quarterly refreshes
+keep it current.
+
+## What This Replaces
+Comparable facility-level FDA regulatory intelligence from incumbent enterprise
+platforms requires annual contracts and enterprise procurement. This dataset
+delivers the same facility-level regulatory history and escalation analytics as a
+one-time, analysis-ready download — no contract, no seat licenses, no platform
+lock-in.
+
+## Source
+All data derives from U.S. FDA public records:
+- **FDA Data Dashboard API** — inspections, inspection citations, inspection classifications,
+  compliance actions (Warning Letters / Injunctions / Seizures), and import refusals.
+- **openFDA** — 510(k) device clearances, establishment registrations, device & drug recall
+  enforcement reports, and device classification.
+- All sources are U.S. Government works in the **public domain (CC0)**. No PII is included
+  (no patient, reporter, or individual-inspector identities).
+
+## The Premium Fields
+The value of this dataset is the **computed regulatory intelligence**, not the raw counts:
+
+- **`facility_risk_score` (0–100)** — a transparent composite of a facility's regulatory exposure.
+  The full formula is published (see Methodology) so you can audit every point or recompute with
+  your own weights from the underlying counts, which are all included.
+- **`ever_escalated_to_wl` / `fastest_escalation_days`** — whether, and how quickly, this facility's
+  inspections have historically escalated to Warning Letters. Directionality is enforced: an action
+  only counts if it post-dates the inspection.
+- **`repeat_citation_flag` / `most_cited_cfr` / `distinct_cfr_codes`** — citation analytics that
+  surface chronic, repeated deficiencies (the same CFR cited across multiple inspections).
+- **Raw counts alongside every signal** — `oai_count`, `warning_letter_count`, `recall_count`,
+  `import_refusal_count`, etc. are all present so you are never locked into our thresholds.
+
+## Methodology
+- **Entity resolution.** Records are joined on the FDA **FEINumber** (Facility Establishment
+  Identifier), normalized to a trimmed string across every source. 510(k) clearances and import
+  refusals carry an FEI directly.
+- **Escalation matching.** For each inspection, the dataset finds compliance actions at the *same
+  FEI* whose action date is **strictly after** the inspection end date, and keeps the **nearest**
+  one (smallest positive day gap). Directionality is enforced in the join itself, so a Warning
+  Letter that predates an inspection can never be counted as escalation. Escalation actions are
+  Warning Letters (the canonical case) plus the stronger judicial actions Injunction and Seizure;
+  the action type is retained so you can restrict to Warning-Letter-only.
+- **Risk score.** Computed from each facility's history with a published formula:
+
+  ```
+  facility_risk_score = min(100,
+      oai_count                              × 15
+    + warning_letter_count                   × 20
+    + recall_count                           × 10
+    + min(import_refusal_count × 2, 20)           (import-refusal contribution capped at 20)
+    + (repeat_citation_flag ? 10 : 0)
+    + (most_recent_classification == OAI ? 15 : 0)
+    + recency_factor)
+
+  recency_factor (awarded only when the most recent inspection outcome was OAI or VAI):
+    +10 if that inspection was within 1 year, +5 if within 3 years, +2 if within 5 years
+    (evaluated as of the 2026-06-23 build date)
+  ```
+
+- **Recall attribution.** Recall enforcement reports carry no FEI, so recalls are matched to
+  facilities by **normalized firm name** (uppercased, punctuation and common legal suffixes
+  removed). The match is deliberately restricted to **unambiguous 1:1 names** — a normalized name
+  owned by exactly one facility — to prevent a multi-site chain's recalls from being falsely
+  attributed to every location that shares its name.
+
+## Validation
+- **Escalation signal validated (reproducible from this file).** Grouping the shipped rows by
+  `most_recent_classification_code` and averaging `ever_escalated_to_wl`, facilities whose most
+  recent inspection was OAI (Official Action Indicated) have escalated to a Warning Letter at
+  **65.7%**, VAI at **6.0%**, and NAI (No Action Indicated) at **2.5%** — a monotone OAI > VAI > NAI
+  ordering (~26× OAI-vs-NAI) confirming the computed escalation fields correlate correctly with
+  inspection severity. (If escalation were a join artifact, NAI and OAI would escalate at similar
+  rates.)
+- **Directionality enforced.** Zero inspections have a negative inspection→action day gap; every
+  Warning Letter counted as escalation post-dates its inspection.
+- **No orphaned escalations.** Every recorded escalation action type exists in the source
+  compliance-action data.
+- **Recall attribution restricted** to unambiguous 1:1 facility-name matches to prevent false
+  attribution (see Methodology).
+- **Completeness.** Row counts reconcile exactly to the validated master tables.
+
+## Known Limitations
+- **Citation coverage.** Inspection citations cover electronically generated, finalized
+  inspections; some manually prepared Form 483s are not present in the FDA citations dataset.
+- **Recall attribution (~35%).** Only recalls with an unambiguous 1:1 firm-name match are
+  attributed; multi-site chains that share a normalized name are deliberately excluded to avoid
+  false attribution, so recall counts are conservative (under- rather than over-counted).
+- **Inspector-level detail not included.** The FDA citations API does not expose individual
+  inspector identity, so inspector-level fields are out of scope for this version.
+- **Risk score is a heuristic.** `facility_risk_score` is a transparent composite indicator, **not**
+  an official FDA designation or prediction. All underlying counts are included so you can recompute
+  it with your own weights.
+- **Warning Letter coverage.** Warning Letters are attributed to facilities that appear in the
+  inspection record; Warning Letters issued to firms without an inspection in the dataset are not
+  rolled into a facility row.
+- **Null-by-design escalation timing.** The `fastest_escalation_days` field is null by design for
+  facilities that never escalated to a Warning Letter (`ever_escalated_to_wl` = false). This covers
+  ~95% of facilities, reflecting that most FDA inspections do not lead to enforcement.
+- **Null-by-design citation field.** The `most_cited_cfr` field is null for facilities with no
+  posted citations on record (~64% of facilities). This is accurate — many inspections,
+  particularly those classified NAI, result in no citations. All citation counts are included so
+  this is transparent.
+
+## Field Definitions
+| Field | Type | Description |
+|-------|------|-------------|
+| `fei_number` | string | FDA Facility Establishment Identifier (FEI) — the cross-source join key. |
+| `legal_name` | string | Facility legal name, taken from its most recent inspection record. |
+| `city` | string | Facility city (most recent inspection record). |
+| `state` | string | Facility state/province (most recent inspection record). |
+| `country` | string | Facility country (most recent inspection record). |
+| `product_types` | list<string> | Distinct FDA product areas inspected at this facility (e.g. Drugs, Devices, Foods). |
+| `total_inspections` | integer | Count of distinct FDA inspections on record for this facility. |
+| `oai_count` | integer | Inspections classified Official Action Indicated (OAI) — the most serious outcome. |
+| `vai_count` | integer | Inspections classified Voluntary Action Indicated (VAI). |
+| `nai_count` | integer | Inspections classified No Action Indicated (NAI). |
+| `most_recent_inspection_date` | date | End date of the facility's most recent inspection. |
+| `most_recent_classification` | string | Classification of the most recent inspection (full text). |
+| `most_recent_classification_code` | string | Classification code of the most recent inspection (NAI/VAI/OAI). |
+| `warning_letter_count` | integer | Count of Warning Letter compliance actions issued to this FEI. |
+| `recall_count` | integer | Count of device + drug recalls attributed to this facility. Recalls carry no FEI, so they are matched by unambiguous 1:1 normalized firm name (see Methodology). |
+| `import_refusal_count` | integer | Count of FDA import refusals recorded against this FEI. |
+| `total_citations` | integer | Total inspection citations (483 observations) recorded across all inspections. |
+| `distinct_cfr_codes` | integer | Number of distinct CFR codes cited across the facility's inspections. |
+| `repeat_citation_flag` | boolean | True if the same CFR code was cited across 2+ distinct inspections. |
+| `most_cited_cfr` | string | The CFR code cited across the most distinct inspections at this facility. Null for facilities with no posted citations on record (~64% of facilities). |
+| `ever_escalated_to_wl` | boolean | True if any inspection at this facility was followed by a Warning Letter (or stronger). |
+| `fastest_escalation_days` | integer | Shortest observed gap (days) from an inspection end to a subsequent Warning Letter; null by design when the facility never escalated (`ever_escalated_to_wl` = false), which covers ~95% of facilities. |
+| `has_510k_clearance` | boolean | True if the facility holds at least one 510(k) device clearance. |
+| `count_510k` | integer | Number of 510(k) device clearances linked to this facility's FEI. |
+| `facility_risk_score` | float | Composite regulatory-risk score, 0–100 (higher = more regulatory exposure). Heuristic, not an official FDA designation. Formula: min(100, oai_count*15 + warning_letter_count*20 + recall_count*10 + min(import_refusal_count*2,20) + repeat_citation_flag*10 + (recent OAI)*15 + recency_factor). |
+
+## Update Cadence
+Quarterly, tracking FDA Data Dashboard refreshes.
+
+## License
+Derived entirely from U.S. federal government public records (FDA Data Dashboard, openFDA), which
+are in the public domain. This structured, enriched version is provided under **CC BY 4.0** —
+attribution required.
+
+## How to Access the Full Dataset
+Full dataset available at [Ruby Intelligence on Gumroad](https://rubyintelligence.gumroad.com/l/izpoe).