US NTSB Aviation Accident and Incident Final Reports Dataset (2016–2023)

This dataset consolidates information from final reports on aviation accidents and incidents occurring between 2016 and 2023, as provided by the U.S. National Transportation Safety Board (NTSB) and available through the NTSB CAROL platform as of December 24, 2024. It covers 7,462 individual occurrences and integrates both structured and unstructured data extracted from the official reports.

The dataset includes detailed data for each occurrence, such as event identification (e.g., NtsbNo, EventID, ReportNo), occurrence details (EventDate, City, State, Country, Latitude, Longitude), aircraft information (Make, Model, AirCraftCategory, NumberOfEngines, EngineType), flight and operator data (Operator, PurposeOfFlight, Scheduled, FAR), injury and damage statistics (FatalInjuryCount, SeriousInjuryCount, MinorInjuryCount, OnGroundInjuryCount, AirCraftDamage), environmental conditions (WeatherCondition), and investigation-related information (ProbableCause, Findings, BroadPhaseofFlight, ReportStatus, ReportUrl, DocketUrl). The dataset also include full text content of the associated reports, extracted from the original PDF files (rep_text).

This dataset is intended to support research on aviation safety, risk analysis, and natural language processing (NLP) applications.