Transparent by design
Restaurant inspection data sources & methodology
Primary inspection source
Restaurant names, CAMIS identifiers, addresses, cuisines, coordinates, inspection dates, actions, violation descriptions, critical flags, scores and letter grades come from the NYC Department of Health and Mental Hygiene’s DOHMH New York City Restaurant Inspection Results dataset.
The source is updated by NYC. Our build downloads a complete snapshot, validates row and restaurant counts, and refuses to replace the existing snapshot when the result is implausibly small.
Restaurant and inspection grouping
Rows are grouped by CAMIS, NYC’s restaurant identifier. Multiple rows can describe separate violations from the same inspection, so inspection events use the synthetic key CAMIS + inspection date + inspection type. Every grouped event must have consistent action, score, grade and grade date; otherwise the build fails closed.
Identical source citations are preserved as source multiplicity but displayed once and never double-counted in violation totals, recurrence or aggregates. The “latest comparable inspection” is the most recent non-ambiguous Cycle or Pre-permit Operational initial/re-inspection with a numeric score. Other event types remain visible in history but do not drive score deltas or peer comparisons.
The displayed letter uses the documented DOHMH selection rules for the most recent selectable grade represented in the snapshot. It is not described as a live posted grade, and we never infer A, B or C merely from a score when the source’s grade-selection rules do not yield one.
Neighborhood source
Neighborhood labels come from NYC Planning’s 2020 Neighborhood Tabulation Areas, version 26b. We assign restaurant coordinates to residential NTA polygons at build time, validate the borough prefix and suppress missing or conflicting assignments. NTAs are statistical areas, not definitive neighborhood boundaries, and can differ from locally recognized names.
Derived comparisons
Borough, cuisine and NTA comparisons use one latest comparable inspection per restaurant and only compare the same inspection program and subtype in the same source snapshot. A comparison is shown only when the cohort contains at least 50 records. “Fewer points than 70%” means 70% of comparable scores are numerically higher; ties are not counted as outperformed.
Neighborhood medians use the same comparable cohort rules. Recurring violation codes count distinct grouped inspection events, never duplicated source rows. These are site calculations, not official NYC ratings.
Publication and quality controls
Map observations can cover the full validated dataset, while indexable restaurant profiles are released as complete geographic clusters. The publication manifest controls profile routes, neighborhood routes, search links, related links and the sitemap. A profile outside the manifest does not exist as an indexable route.
Limitations
- Inspection data is a historical snapshot and conditions can change after an inspector leaves.
- A low score does not measure food quality, service, value or popularity.
- Restaurant names, cuisine labels, addresses and phone numbers reflect the official source and may lag business changes.
- Percentiles describe the source snapshot and are not an official NYC metric.
- This site is not a substitute for official public-health guidance.
Update timestamps
Source rows: 295,256. Distinct CAMIS IDs in raw snapshot: 31,274. Downloaded August 15, 2026 at 12:04 AM UTC.