The Oil & Gas Refining & Marketing Data Guide (2026)

The U.S. Energy Information Administration anchors the field with three records. The Petroleum & Other Liquids Data Portal is the only quality-10 record here, with daily spot prices back to 1986, PADD/state/refinery granularity and a PET bulk file covering roughly 53,300 series; the annual Refinery Capacity Report and the Short-Term Energy Outlook both score 9. Government statistical offices supply much of the rest: Eurostat publishes monthly stock tables for 41 geographies, the European Commission's Weekly Oil Bulletin tracks pump prices across the EU and UK, the Canada Energy Regulator ships pipeline throughput CSVs, India's PPAC issues its Monthly Ready Reckoner, and data.gov exposes a discovery layer of about 400 petroleum-keyword federal records.

Industry bodies run their own transparency platforms. Gas Infrastructure Europe operates AGSI+ and ALSI, both free JSON APIs that require registration and source attribution. NGOs and multilaterals round out the set: Global Energy Monitor's Global Oil Infrastructure Tracker holds asset-level records for 1,634 oil pipelines totalling about 487,000 km, KAPSARC runs a roughly 1,300-dataset Gulf-weighted portal, and IEA Monthly Statistics is republished on energydata.info under non-open terms.

The 16 primary records sort into four working groups: US federal anchors, European official statistics, regulator files and infrastructure trackers, and the transparency-platform tier topped off with one commercial outlier.

US federal anchors

European official statistics

• EU Weekly Oil Bulletin (quality 8). The European Commission's reference cross-country pump-price series: weekly consumer prices for euro-super 95 petrol, automotive diesel, heating oil, fuel oil and LPG in every EU member state plus the UK, published with and without taxes. The history workbook is about 4.24 MB with 7 sheets, roughly 1,100 weekly rows and 226 columns in the prices-with-taxes sheet, running weekly from 5 January 2003. Reproduction is authorised provided the source is acknowledged, though the Commission cautions that cross-country comparisons are of limited validity because of differences in product quality and market structures.
• EU Weekly Oil Bulletin - cleaned dataset, Hugging Face (quality 7). A community-cleaned, tidy rebuild: one row per country, week and fuel type with pump prices in EUR per litre, weekly from 3 January 2005 through 22 June 2026 across the 27 EU member states plus the UK. The parquet build holds 135,867 rows (the dataset card cites 106,028) and ships under commercial delivery terms 4.0 in Parquet, CSV and JSON.
• Eurostat Stock Levels for Gas Products, nrg_stk_gasm (quality 9). Monthly opening and closing stock levels for 41 geographies - EU-27, the euro area, member states and neighbouring countries including Norway, the UK and Ukraine - from January 2008, roughly 38,800 observations across 223 months, delivered as JSON-stat and SDMX-ML under commercial delivery terms 4.0. The full JSON-stat extract is about 640 KB.

Regulator files and infrastructure trackers

Transparency platforms, regional portals and the commercial outlier

Ten recurring jobs define how teams work this pool, and each maps to named sources:

Datadory delivers this dataset fully structured and normalized, ready for direct analysis without manual pipeline configuration or schema parsing.

Four personas dominate demand for downstream oil data, and each searches for it differently.

Market Researchers & Consultants. They type "us petroleum market statistics by state," "petrol prices by european country comparison weekly" and "gulf middle east energy market statistics." This is the best-covered persona in the slice: all 12 relevant datasets rate strongly, anchored by citable EIA tables, the Commission bulletin and KAPSARC's Gulf depth. Across the catalog, 786 datasets are strongly relevant for this persona and 665 of those (84.6%) are completely free. See the market researchers hub for the ranking.

Journalists, Academics & Students. Coverage runs deep - 13 relevant datasets with 11 rated strongly - for prompts like "how many us refineries operating capacity official," "official eu weekly petrol price bulletin" and "europe petrol prices historical dataset for research." Public-domain EIA reports, the Commission's weekly bulletin and commercial delivery terms 4.0 builds resolve most citation needs without permissions letters.

A working downstream stack can be assembled in layers, entirely on free sources.

Layer two: add the European panel. Load the Hugging Face cleaned bulletin parquet - one row per country, week and fuel, 135,867 rows under commercial delivery terms 4.0 - instead of parsing the Commission's 4.24 MB workbook, and keep the raw Weekly Oil Bulletin for pre-tax versus post-tax splits. Query Eurostat's nrg_stk_gasm as JSON-stat for harmonized stocks, and register once for the AGSI+ and ALSI keys, hard-coding the required "Source: GIE" credits into every export.

Every figure below is computed from Datadory's catalog snapshot for this industry, as of August 2026.

Where can I get free historical gasoline and diesel price data?

The EIA Petroleum & Other Liquids Data Portal is the default answer. It publishes daily spot prices from 1986, weekly status series from the early 1990s and monthly marketing series back to the mid-1970s, including the pswrgvwall.xls gasoline history file. Everything is US commercial delivery terms, and the PET bulk file alone covers roughly 53,300 series.**

Is refinery-level capacity data publicly available in the United States?

Yes. The EIA Refinery Capacity Report, an annual Form EIA-820 survey taken as of January 1, publishes atmospheric crude distillation, downstream charge and production capacities for about 130 respondent sites. Current editions ship as refcap26.xlsx with roughly 3,336 data rows, and historical reports reach back to 1994.**

How do I compare fuel prices across European countries?

Use the European Commission's Weekly Oil Bulletin, which reports weekly consumer prices for euro-super 95 petrol, automotive diesel, heating oil, fuel oil and LPG in every EU member state plus the UK, with and without taxes. The history workbook covers weekly data from January 2003, and a commercial delivery terms 4.0 cleaned version exists on Hugging Face.**

Which datasets show how full European gas storage is?

What is the best machine learning dataset for fuel price forecasting?

For a clean supervised panel, the Hugging Face 'EU Weekly Oil Bulletin - cleaned dataset' is purpose-built: one row per country, week and fuel type with pump prices in EUR per litre, weekly from January 2005 through June 2026. The parquet build holds 135,867 rows under commercial delivery terms 4.0, so training-permissive reuse needs only attribution.**

Where can I find India's refinery throughput and petroleum product data?

India's Petroleum Planning & Analysis Cell publishes it free each month. The Ready Reckoner pairs a roughly 45-page PDF with an Excel workbook of 107 worksheets covering installed refinery capacity and crude processing by company, product-wise production and consumption, imports and exports, prices, taxes and gross refining margins, with fiscal-year series from roughly FY2012-13.**

Are station-level gas prices available through an API?