Look up the fatal-crash history near any US address — free, no signup. Built on 254,000+ NHTSA FARS crashes (2017-2023), spatially joined to FHWA HPMS traffic volume so you get the per-VMT rate, not just the count. The raw FARS data is public and free from NHTSA; what we add is the joining, decoding, geo-indexing, and a citation-ready baseline.
FARS API (farsapi.com) is a REST API for NHTSA's Fatality Analysis Reporting System, covering 254,000+ US fatal crashes from 2017-2023. Each crash is pre-joined to FHWA HPMS traffic volume so queries return both the raw count and the AADT exposure denominator. Responses are JSON; radius search, intersection lookup, and aggregate statistics are built in. Free tier for statistics, $19/month for full access. The underlying FARS dataset is public domain and free to download directly from NHTSA — what you pay for here is the pre-joining, label decoding, geo-indexing, and traffic-volume baseline, not the raw records.
Raw fatal crash counts are misleading without traffic volume. A busy interstate will always have more crashes than a quiet residential street. To compare locations honestly you need the per-VMT rate, and to compute that you need the AADT for the road segment each crash happened on. NHTSA's FARS doesn't include AADT. FHWA's HPMS does, but it's a separate dataset in a separate format. We join the two and persist the result, so every location query returns the rate, not just the count.
17 observed vs 0.5 expected
For an Interstate at this AADT (70,630/day average) over 6 years, the national baseline rate predicts ~0.5 fatal crashes. The Connector had 17. That's the kind of statement that goes in a brief, an actuarial memo, or a city safety report.
Each of the 254,877 FARS crashes 2017-2023 is snapped to its nearest FHWA HPMS road segment within an 80m radius. 80% of crashes match. The matched segment carries AADT, road class (Interstate, Principal Arterial, etc), lane count, and speed limit. Per-road-class national baselines computed from the joined population: Interstates 0.32 fatal crashes per 100M VMT, Major Collectors 1.89. Methodology →
NHTSA publishes 20+ CSV tables with numeric codes that change between years. An API that uses make=12 instead of "Ford." No location search. No exposure adjustment. We pre-join everything, decode labels, normalize across years, and add geo queries. 254K crashes, 392K vehicles, 625K persons.
$9 per report. Map, charts, AI-written analysis, every crash record with vehicles + persons, AADT-adjusted Expected vs Actual, segment-level disambiguation, formal NHTSA citation, print-ready PDF. Used by plaintiff attorneys for discovery and insurance teams for risk assessment.
Traffic accident data API for insurance actuaries. Vehicle risk profiles, geographic crash density, DUI accident statistics by state. Historical accident data for insurance pricing models. Road accident data API by state with aggregate breakdowns.
Crash history at intersection API for legal discovery. "Show me every fatal crash within 500 feet of this intersection." Intersection collision data with full vehicle and person detail. $9 per report, no subscription.
Truck accident data API filtered by body type. Rollover rates by road class. Drunk driving accident data and time-of-day risk profiles. Traffic collision data API for risk assessment and benchmarking against national baselines.
Vehicle accident data API for developers building safety tools, location intelligence, or risk scoring. Free statistics tier. FARS data access methods via clean REST API. No more downloading CSVs and joining 20 tables in R or Python.
| Feature | FARS API | NHTSA API | Kaggle | CSV |
|---|---|---|---|---|
| Traffic volume (AADT) joined to crashes | Yes | No | No | No |
| Expected vs Actual fatal crashes | Yes | No | No | No |
| Radius / geo search | Yes | No | No | No |
| Intersection lookup | Yes | No | No | No |
| Pre-joined tables | Yes | Partial | No | No |
| Human-readable labels | Yes | Codes | Varies | Codes |
| Aggregate stats | Yes | No | No | No |
| YoY normalization | Yes | No | No | No |
| Documentation | Full | Poor | Min | Min |
Enter coordinates or click a sample corridor below.
Detailed crash statistics for every US state - counties, cities, time patterns, vehicles, and conditions.
Every fatal motor vehicle crash in the US, 2017–2023. From NHTSA FARS, the definitive US fatal crash dataset.
Refreshed annually when NHTSA publishes. We handle the schema changes, code normalization, and table joins.