Open Data Downloads

The datasets behind VinSpecs, as plain CSV — 312 rows across 4 files, each with its source, licence and verification date.

Every dataset VinSpecs publishes is downloadable here as a plain CSV — 312 rows across 4 files. Each row of each file traces to the source named below it (official DMV schedules, EPA ratings, real classifieds listings), the licence is stated, and nothing is hand-typed: every file is a static export of a snapshot the site already serves.

Used-value retention by model line

230 rows · columns: make, model, year, original MSRP, median used asking price, retained %, listing sample

Download used-value-retention.csv ↓

State title fees (all 50 states)

50 rows · columns: state, state code, title fee USD, notes, verified-official flag, verification date, official source URL

Download us-state-title-fees.csv ↓

  • Source: State DMV official fee schedules, one URL per row
  • Licence: Public government fee schedules; VinSpecs compilation
  • Used on: /states

Promised range vs. EPA rating (20 landmark EV model years)

20 rows · columns: vehicle, trim, promised model year, promised range mi, EPA rated mi, difference %, verdict

Download ev-promised-vs-epa.csv ↓

EV efficiency by vehicle class (1,197 BEVs aggregated)

12 rows · columns: vehicle class, vehicles measured, Wh/mi median, min, max, latest model year

Download ev-efficiency-by-class.csv ↓

How these files are produced

Each CSV is an export of a JSON snapshot that the live pages already compute from — the same numbers you see on the page are the numbers in the file, and every file was generated on 2026-11-18. The exports are refreshed with each data ingest (the ingestion runs on a schedule through the hosting cron), so a file's generation date tells you exactly which snapshot vintage you are looking at. If you find a discrepancy between a CSV and the page it feeds, that is a bug worth reporting: corrections policy.

What is deliberately not in these files

These are the datasets whose full contents can be published: aggregations of public listings and official fee schedules. Individual transaction prices, VIN-specific histories and anything identifying a seller are not in the snapshots and never will be. The full inventory of every source the site uses — nine data layers with licence names — lives at /data-sources, and the methodology notes for the value-retention file are in the retention study.

All data sources → Value-retention study →