Spain Vehicle Fleet by Brand, Model & Province (DGT)

Query Spain's annual parque de vehículos from the Dirección General de Tráfico (DGT) as compact cohort data: province × brand × model × fuel × category × year. The Actor returns vehicle counts, not individual records. It is useful for dealerships, insurers, EV analysis, market studies, fleet research and agents connected through MCP.

The annual aggregate is bundled in the Actor image. Runtime requests do not download the DGT's approximately 1.6 GB source ZIP, so filtered runs are fast and predictable. The Actor is independent and not affiliated with the DGT.

Source and coverage

• DGT annual consolidated fleet microdata, 2025.
• Source ZIP: https://www.dgt.es/microdatos/Parque/2025/parque_consolidado_2025_PROVINCIA.zip
• 52 province files, pipe-delimited TXT, CP1252, 45 source columns.
• Cache metadata records the source URL, retrieval headers, ETag and Last-Modified value.
• Aggregated output contains no registration plates, vehicle identifiers or VINs.

Official fuel mapping (PROPULSION)

The mapping follows the DGT's annual interface dictionary, not an empirical inference. Official PDF: https://www.dgt.es/export/sites/web-DGT/.galleries/downloads/dgt-en-cifras/publicaciones/parque-de-vehiculos/Interfaz-de-Salida-Fichero-Parque-Anual.pdf

Code	Fuel
0, G	Gasolina / Petrol
1, D	Diésel / Diesel
2	Eléctrico / Electric
3	Otros / Other
4	Butano / Butane
5	Solar / Solar
6	GLP / LPG
7	GNC / CNG
8	GNL / LNG
9	Hidrógeno / Hydrogen
A	Biometano / Biomethane
B	Etanol / Ethanol
C	Biodiésel / Biodiesel
empty	No consta / Not reported
undocumented code	Desconocido / Unknown

Input

{
  "province": "28",
  "marca": "",
  "combustible": "",
  "categoria": "",
  "year": 2025,
  "maxResults": 1000,
  "exportFiles": ["csv", "xlsx"],
  "exportLanguage": "es"
}

province accepts a code or name such as 28 or Madrid. Brand, fuel and category filters are case- and accent-insensitive. Results are sorted by vehiculos descending. maxResults defaults to 1,000; set it to 0 for all matching cohorts. The first 10,000 returned cohorts are chargeable; more rows are still returned without an additional own charge.

Output

Each dataset item is one cohort:

{
  "provincia": "28",
  "provinciaNombre": "MADRID",
  "marca": "VOLKSWAGEN",
  "modelo": "GOLF",
  "combustible": "Diésel",
  "combustibleCodigo": "1",
  "categoria": "TURISMOS",
  "anio": 2025,
  "vehiculos": 12345
}

SUMMARY contains separate scopes for the complete filtered population and the exact rows returned after maxResults. It also contains SUMMARY.billing: cohortsReturned, cohortsCharged, capApplied, userLimitReached, the nominal cost, and the count of vehicles carrying undocumented fuel codes.

CSV and XLSX contain exactly the rows delivered to the dataset. XLSX also includes a provenance sheet; the data sheet has a frozen header and autofilter.

Pricing

There is one own charge event: cohorte-charged at USD 0.001 per returned cohort. The platform handles its synthetic actor-start event. This Actor does not charge a report or a separate start event. The own event cap is 10,000 per run.

Real nominal examples from the regenerated 2025 cache:

Query	Cohorts returned	Maximum own charge
Small query, 20 cohorts	20	about USD 0.02
Madrid without a result limit	24,512	USD 10.00
Spain without a result limit	375,852	USD 10.00

The CLI smoke runner reports nominal event cost; Apify platform compute cost is separate.

Examples

• examples/input.json — small general query.
• examples/input-madrid-diesel.json — Madrid diesel passenger vehicles.
• examples/input-national.json — complete national cohort export.

FAQ

Which year is covered? Version 1 contains the annual 2025 fleet file. Unsupported years are rejected; they are never silently substituted.

Are plates or VINs included? No. Results are aggregate cohorts and contain no plates, VINs or individual vehicle identifiers.

Is this an official DGT product? No. It transforms an official open-data source but is not operated by, endorsed by or affiliated with the DGT.

Can an agent use it through MCP? Yes. Its simple JSON input and dataset/CSV/XLSX outputs are suitable for MCP-connected research and workflow agents.

Development

Internal column notes, cache regeneration, provenance, fuel-dictionary details and test commands are in docs/DEVELOPMENT.md.