The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.

NHTSA ODI investigations (5,334 pre-recall safety probes)

Every NHTSA Office of Defects Investigation action since the agency began posting them — Preliminary Evaluations (2,326), Engineering Analyses (1,930), Defect Petitions (502), Recall Queries (244), Audit Queries, Tire Investigations. These are the agency probes that often lead to recalls: Tesla autopilot investigations, Takata airbag EA, Ford fuel pump EA, GM ignition switch EA. Each record has the action number (AQ/PE/EA/RQ-prefix), open/close dates, affected makes + components + model years, manufacturer, subject, and summary text. Source: static.nhtsa.gov/odi/ffdd/inv/FLAT_INV.zip (4MB zipped → 389MB raw tab-delim). Distinct from nhtsa_recalls (formal recalls) and nhtsa_complaints (consumer reports).

Live API

This dataset is served via a live REST API at api.ai-analytics.org. The card you're reading exists so HuggingFace's index can route AI agents + researchers to the canonical source.

Quick start

# Recent records (JSON)
curl 'https://api.ai-analytics.org/api/v1/nhtsa/investigations/recent?limit=10'

# Browse on the web
open 'https://api.ai-analytics.org/datasets/nhtsa-investigations'

MCP (Model Context Protocol)

This dataset is also accessible via MCP at https://api.ai-analytics.org/mcp — Claude, ChatGPT, and other MCP-compatible agents can call it directly.

Source

Keywords

NHTSA, ODI, defect investigation, preliminary evaluation, engineering analysis, defect petition, Takata, Tesla autopilot, recall query

Citation

@misc{ai_analytics_nhtsa_investigations_2026,
  title  = {{name}},
  author = {AI Analytics},
  year   = {2026},
  url    = {https://api.ai-analytics.org/datasets/nhtsa-investigations},
  note   = {Derived from US National Highway Traffic Safety Administration. License: CC0 1.0.}
}

Related


Maintained by AI Analytics. Data is CC0; please cite the upstream agency.

Downloads last month
9