Datadory notebook
Epa Fuel Economy Dataset Csv Download Data: Dataset Structure and Field Coverage
1,744 datasets. Pick your catch. Every guide here is built on what the catalog can actually prove.
1,744 datasets. Pick your catch.
What does the combined 1984-2026 CSV actually contain?
Each row is one vehicle configuration - a specific combination of model year, make, model and engine/transmission option - tested at the EPA National Vehicle and Fuel Emissions Laboratory in Ann Arbor, Michigan, or reported by manufacturers under EPA oversight. The combined file spans 43 model years, 1984 through 2026; Datadory's review notes tens of thousands of configuration records, anchored on per-year ZIP archives of roughly 1 MB each, since the publisher states no row count on the page.
Which EPA fuel economy file should you use for what? (ranked)
Five artifacts come out of this source family, and picking wrong wastes hours. Ranked by how often each answers a real question:
How do you load and validate the file in practice?
A four-step walkthrough that holds up in a notebook or a scheduled job:
What else pairs well with the EPA fuel economy data?
Three complements cover what a ratings table cannot. For fleet stock rather than rated efficiency, Washington State Electric Vehicle Population Data publishes 294,193 registered BEV/PHEV records refreshed weekly under commercial delivery terms, geocoded from county down to census tract - Datadory maintains a head-to-head comparison of that registry against this EPA file. For identity resolution, NHTSA vPIC decodes VINs into up to 154 fields with no integration key and scores 10/10, the slice's highest. For Europe, ACEA's monthly registration releases and Eurostat's commercial delivery terms 4.0 STS_INPR_M production index cover demand and output across 42 reporting areas.
Historical tabular context also exists inside the catalog: the UCI Automobile Dataset offers 205 rows of 1985 cars under commercial delivery terms 4.0 - useful as a teaching benchmark, tiny next to 43 model years of EPA records. And if your analysis extends to production volumes, OICA's country-by-country tables run from 1999 through 2026.
Who builds on this dataset?
Six of Datadory's eight personas rate the record relevant, three at maximum relevance. Data scientists use it as a benchmark tabular set - every 1984-2026 model year at configuration level with MPG, CO2 and drivetrain fields. Market researchers treat it as an authoritative 40+ year panel of US specifications and efficiency by maker. Journalists and academics cite it as the standard public-domain reference for US fuel-economy reporting, detailed on our journalists academics use cases page. Competitive-intel teams track manufacturers' MPG/CO2 trajectories year over year, investors read long-run efficiency trends as compliance-risk evidence, and developers stand up vehicle-spec lookup services on top of the free bulk files - the workflow our developers builders use cases page walks through.
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
EPA Fuel Economy Dataset (1984-2026)
EPA FuelEconomy.gov Web Services API Data
Washington State Electric Vehicle Population Data
NHTSA vPIC Vehicle Manufacturer and VIN API Data
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sampleQuestions worth asking
How far back does EPA fuel economy data go?
The combined CSV covers model years 1984 through 2026 - 43 model years in one file. Per-model-year packages reach back further, to 1978, and already include a preliminary 2027 entry alongside the printed Fuel Economy Guide PDFs.
Should I use the EPA CSV or the fueleconomy.gov API?
Use the bulk CSV when you need the whole panel: analysis, modeling, manufacturer trend work. Use the EPA FuelEconomy.gov Web Services API - quality score 8/10, keyless JSON or XML - when an application needs single-vehicle lookups by year, make and model. Both describe the same universe of configurations.