Spot vehicle defects building up, week by week.
We turn NHTSA's consumer complaints and official recall file into weekly signals by automaker, component and stock ticker, scored against each one's past 52 weeks.
This week's largest spike
Data refreshed daily at 18:00 UTC from NHTSA.
This week's largest spikes
Complaints filed with NHTSA in the latest complete week, compared with each make and component's own past year.
| Make | Component | Complaints | Typical week | Surprise score | Elevated (last 3 wks) | Crashes |
|---|---|---|---|---|---|---|
| Loading… | ||||||
A surprise score of 2.5 or more happens in about 0.6% of normal weeks. 2 or more is worth watching.
How the score works
Collect
Every day we pull new complaints from NHTSA and its complete official recall file. Each record is stamped with the time we first saw it, so you can backtest without hindsight.
Score
Complaints are counted per make, component and week, then compared with the previous 52 weeks using a model built for small, noisy counts. The current unfinished week is never scored.
Deliver
A REST API maps every make to its listed parent and the ticker valid at the time, so Jeep before 2021 is FCAU and after is STLA. Results page cleanly and errors are consistent.
Does it work? The backtest
We took every week since 2014 in which a make's component was scored, skipped weeks right after a recall of that same component (recalls cause complaints), and checked how often a new recall of that component followed within 26 weeks.
| Surprise score of at least | Spikes | Followed by a recall | Compared with a normal week |
|---|---|---|---|
| Loading… | |||
Most spikes are not followed by a recall. Use this as an early-warning input alongside your own research, not as a prediction.
Who uses it
Investment research
Track defect pressure on Ford, GM, Stellantis, Tesla, Toyota, Honda, Hyundai and Kia with point-in-time data that is safe to backtest.
Parts and e-commerce
See which components are drawing complaints for which models before replacement-part demand shows up in your sales.
Insurance, fleets and safety
Monitor the vehicles you cover or run, and get the recall record and safety advisories in one place.
Start in five minutes
Subscribe on Rapid, copy your key, and call the API. Rapid shows a ready-made snippet for your language with the exact host.
import requests
r = requests.get(
"https://YOUR-RAPID-HOST/v1/signals",
params={"ticker": "F", "component": "ALL", "limit": 12},
headers={"X-RapidAPI-Key": "YOUR_KEY", "X-RapidAPI-Host": "YOUR-RAPID-HOST"},
)
for week in r.json()["data"]:
print(week["week_start"], week["complaints"], week["surprise_z"])| Endpoint | What you get |
|---|---|
/v1/signals/latest | The latest week's largest spikes, most surprising first |
/v1/signals | Weekly history for one make or ticker, any component, back to 2014 |
/v1/recalls | Vehicle recalls with affected units, dates, advisories and part numbers |
/v1/complaints | Complaint records by make, model, year and component (no VINs or narratives) |
/v1/entities | Make to parent company to ticker, with the dates each mapping applies |
/v1/methodology | Method, data dictionary and the latest backtest |
Pricing
Monthly plans billed through Rapid. Start free, upgrade or cancel from your Rapid account at any time.
| Plan | Price per month | Requests per month | Good for |
|---|---|---|---|
| Basic | Free | 100 | Trying the data |
| Pro | $29 | 10,000 | Dashboards and side projects |
| Ultra | $99 | 100,000 | Apps and research teams |
| Mega | $299 | 1,000,000 | Full-history pulls and heavy use |
Free sample
A year of weekly parent-company signals and a year of recalls, plus the method and data dictionary. No sign-up needed.
- signals_weekly_sample.csvWeekly signals, parent companies
- recalls_sample.csvVehicle recalls, last 12 months
- README_sample.mdMethod, backtest, data dictionary
Questions
Where does the data come from?
From NHTSA's Office of Defects Investigation: the consumer complaints database and the official recalls file. Both are US public domain. We are not affiliated with or endorsed by NHTSA.
How fresh is it?
We check NHTSA every day at 18:00 UTC and rebuild the signals right after. Every response says which date the data is as of.
What does point-in-time mean here?
Every complaint and recall carries the time we first saw it. History from before we started collecting is clearly flagged as estimated, so your backtests can tell the difference.
Which vehicles are covered?
Model years 2015 onward for Ford, Chevrolet, GMC, Cadillac, Buick, Ram, Jeep, Dodge, Chrysler, Tesla, Toyota, Lexus, Honda, Acura, Hyundai and Kia. Equipment-only recalls, such as aftermarket tires, are excluded.
Do you include VINs or complaint text?
No. We keep counts, dates, components and reported crashes, fires and injuries, and leave out VINs and the free-text narratives.
Is this investment advice?
No. These are statistical indicators built from public data. Most spikes are not followed by a recall, so treat them as one input among many.