Methodology

The directory is built from publicly available sources — 2M+ parts classified and mapped. The primary inputs:

• FAA PMA database. Parts Manufacturer Approval records — approved part numbers, approval holders, and applicable products.
• FAA STC records. Supplemental Type Certificates and their associated part references.
• FAA TCDS. Type Certificate Data Sheets for airframe and engine type reference data.
• FAA Service Difficulty Reports (SDR). Field-reported component difficulties, 1995 to present — the ground-truth signal behind reliability and disposition metrics.
• FAA Aircraft Registry (MASTER / DEREG). US-registered aircraft and deregistrations — fleet size, age, operators, and fleet-outflow signals.
• Federal Register — FAA Airworthiness Directives. Published FAA ADs, with part-number references extracted where the AD text permits.
• EASA Airworthiness Directives. European ADs, cross-referenced against FAA ADs where both agencies act.
• BTS T-100 segment data. US DOT Bureau of Transportation Statistics — measured departures and airborne hours, used for fleet utilization.
• BTS Form 41 financials. Airline operating economics from US DOT filings.
• SEC EDGAR filings. Public-company disclosures relevant to fleets, retirements, and the aftermarket.
• Live aircraft activity (adsb.lol). Community-collected ADS-B data used to observe fleet activity. See the attribution note below.
• Major distributor catalogs. Publicly accessible product listings and part descriptions.

Each part is classified using AI-assisted analysis to an ATA chapter, airframe applicability, a capability category, and a part class (rotable / expendable / consumable). Classifications are cross-validated against ground-truth government data — for example, the ATA chapters implied by a part's FAA Service Difficulty Report history are audited against our assigned chapter, and disagreements are reviewed and corrected.

Confidence is tracked per classification, and every classification is logged for audit. Parts that cannot be confidently classified are left honestly unclassifiedrather than guessed — an "unknown" in this database means the evidence wasn't there, not that we didn't look.

Derived metrics are only useful if you know exactly what they measure. Our rules:

• Estimates are labeled as estimates. Measured values (e.g. utilization from BTS T-100 segments) are distinguished from derived ones (e.g. check cadence = published intervals × measured utilization).
• Derived metrics ship with their inputs. Scrap propensity is shown with the underlying SDR replace/repair counts and a confidence band, and rows that rely on a fallback signal are flagged as such.
• Demand signals are not failure probabilities. SDR-based field-demand metrics are directional intensity signals and are never framed as a probability that a part will fail.
• Fleet coverage caveats are stated.Fleet metrics focus on US-registered aircraft (FAA registry). Deregistration is reported as "retirement or export," not assumed to be retirement. And absence of observed activity is never asserted as idleness — coverage gaps exist, and we treat them as gaps.

If you believe a part is mapped to the wrong ATA chapter, airframe, or manufacturer, tell us: info@aeropartsintel.com. Include the part number and what you believe is wrong. Verified corrections are applied to the database.

How uploaded data is handled is covered on the Security page; legal terms are in the Terms of Service and Privacy Policy.