MarvinBregiosa/us-truck-fatalities-fars-2018-2024

US Truck Fatalities (FARS) — 2018–2024

Row-level dataset of every fatal motor-vehicle crash in the United States, 2018–2024, where at least one vehicle was a medium or heavy truck (FARS BODY_TYP codes 60–69). Filtered from NHTSA's Fatality Analysis Reporting System (FARS) National CSV downloads.

Published interactive report based on this data: 👉 [Vision Zero Report Card 2026 — accidentlawyerreview.com](https://accidentlawyerreview.com/research/vision-zero-report-card/)

Reproducible Python pipeline (the join script that produced this CSV from raw FARS zips): [github.com/MarvinBregiosa/vision-zero-fars](https://github.com/MarvinBregiosa/vision-zero-fars)

Related publications:

• —Build notes + bugs (Astro + Python): dev.to write-up
• —Policy framing: Vision Zero is failing in 8 of 17 cities, on Medium
• —On DOIs for indie data: From CSV to citable, on Hashnode
• —Permanent DOI: 10.5281/zenodo.20487070 — Zenodo archive of this same dataset.

What's in the dataset

truck_fatal_crashes_2018_2024.csv — 33,898 rows, 9 columns:

Definition of "truck-involved"

Medium and heavy commercial trucks per FARS BODY_TYP coding:

Passenger pickups under 10,000 lb GVWR are excluded — that's the standard NHTSA / IIHS truck-crash definition.

Headline finding from the report

Of 17 US cities that formally adopted Vision Zero, only 10 cut truck-involved fatalities from a 2018–2019 baseline to the 2023–2024 average. 8 saw them rise. New York City — the first major US Vision Zero adopter (2014) — saw truck deaths rise +13.1% over the same period.

The full ranked report with sortable tables, per-city sparklines, and methodology lives at [accidentlawyerreview.com/research/vision-zero-report-card](https://accidentlawyerreview.com/research/vision-zero-report-card/).

Reproducing the filtered dataset from raw FARS

git clone https://github.com/MarvinBregiosa/vision-zero-fars
cd vision-zero-fars

mkdir -p data
for yr in 2018 2019 2020 2021 2022 2023 2024; do
  curl -o "data/FARS${yr}NationalCSV.zip" \
    "https://static.nhtsa.gov/nhtsa/downloads/FARS/${yr}/National/FARS${yr}NationalCSV.zip"
  unzip -d "data/" "data/FARS${yr}NationalCSV.zip"
done

python3 analyze.py

Caveats before you build models on this

• —Small annual counts at the city level are noisy. Most cities have single-digit truck-fatality counts per year. Aggregate across years before drawing conclusions.
• —City boundaries vary. FARS uses FIPS city codes; through-route crashes at city limits sometimes attribute to the surrounding county instead of the city, so urban truck-deaths are likely undercounted for cities with major interstate-through-routes.
• —2024 is the FARS Annual Report File (preliminary), released April 2026. The Final File will replace it in 2027 and may revise individual counts slightly.
• —Geocoding quality varies. Lat/lon is populated for ~98% of crashes but precision varies by state (some rural states snap to road midpoints).

License & attribution

Released under CC BY 4.0. If you use the data, please cite:

AccidentLawyerReview Editorial Team. Vision Zero Report Card 2026: Truck Deaths in 19 Major US Cities. Published 27 May 2026. <https://accidentlawyerreview.com/research/vision-zero-report-card/>

About

AccidentLawyerReview is an independent public-data directory of US personal-injury law firms. The Crash Atlas series — including this Vision Zero Report Card — is a sister data-research project published alongside the directory. Not a law firm. Not legal advice.