Eurostat Railway Freight Transport Statistics

The official count of what Europe puts on rails. Eurostat's railway freight statistics report the transport performance of the EU rail network from two angles at once: tonnage lifted and tonne-kilometres performed, broken out by reporting country, NST 2007 commodity group, whether the haul was national, international or in transit, dangerous-goods class and consignment type. The headline number for 2024 is 375 billion tonne-kilometres, down 0.8 percent on the year and still chasing the 410 billion tkm peak the network set back in 2018.

The country picture is lopsided in a way analysts can actually use: Germany alone accounts for 126 billion tkm, roughly a third of the EU total, followed by Poland at 57 billion and France at 32 billion. Tonnage declines are tabulated for all 24 reporting countries, and quarterly series track the year-on-year movement through 2023 and 2024 - so both the trend and the turning points sit in one place.

Everything below describes what Datadory delivers from that catalogue: cleaned, typed, and cut to your countries and commodities before anything ships.

Rows from the August 2026 research pass, exactly as they land before normalization:

# rail_go_grpgood - goods transported by NST 2007 group of goods
# country-year grid, annual series 2014-2024
geo                     : DE              # Germany, largest national market
time                    : 2024
unit                    : MIO_TKM         # million tonne-kilometres
commodity idx 3         : 13254           # million tkm
commodity idx 7         : 12018
commodity idx 6         : 4851

# rail_go_quartal - quarterly goods transported
period                  : 2023Q1 .. 2024Q4   # four quarters x two years

# rail_go_typepas - detailed-reporting goods by type of transport
dimensions              : freq, unit, tra_cov, geo, time

# companion families, same dimensional grammar
rail_go_dnggood         : dangerous goods, 14 classes
rail_go_consgmt         : full train vs wagon load
rail_pa_total           : EU27_2020, 2023, MIO_PKM = 429575 million passenger-km

Three readings fall straight out of those lines. First, concentration: three commodity groups put Germany at 13,254, 12,018 and 4,851 million tonne-kilometres respectively in a single year - bulk flows, not parcel traffic. Second, the quarterly family gives you the same quantities at business-cycle resolution instead of waiting out whole years. Third, the dimensional grammar repeats across every table in the family, so a model built on one table generalizes to the others without re-parsing.

Nine constructs carry the entire family; they recur across the goods tables and their passenger, rolling-stock and totals companions, so one dictionary parses the lot. The verified core is below.

• Geography: the EU27 aggregate plus the wider family of EU, EFTA and candidate countries. Two gaps worth knowing before you plan around them: Cyprus, Malta and Iceland have no railways to measure, and Belgium is excluded from some freight tables for confidentiality reasons.
• Temporal: annual series run roughly 2014 through 2024, with a quarterly layer covering Q1 2023 to Q4 2024 - enough history for structural work, enough recency for current-state questions.
• Granularity: country-year rows on the annual tables and country-quarter rows on the quarterly table, each disaggregated by commodity group, transport coverage, dangerous-goods class or consignment type. Aggregates throughout - never waybill-level, never wagon-level.

Set against Datadory's catalog of 1,744 datasets across 159 industries, this record scores 9/10 for quality - verified field definitions, worked numeric samples and confirmed coverage on the research pass.

API, files, or your warehouse. Daily, weekly, or hourly.

You pick the channel and the cadence; the dictionary above travels unchanged through all three. Analysts modelling European logistics corridors tend toward warehouse load joined to their own geography tables; consultancies benchmarking a client's home market take file extracts cut to named countries and commodity groups; product teams wiring freight-intelligence features take a scoped feed of the markets they serve. Name the countries, commodities and coverage splits when you request the sample - the sample ships first either way, and changing cadence afterward is a settings conversation, not a re-integration project.

• Modal-shift and decarbonisation tracking - tonne-kilometres against road-side equivalents is the honest test of whether freight really moves off roads, and the national/international/transit split shows whose corridors changed rather than just how much.
• Commodity-flow and logistics-site planning - the NST 2007 grid says which goods move where at country-year resolution, which is the demand signal behind terminal, siding and intermodal-hub siting decisions.
• Dangerous-goods exposure analysis - fourteen hazard classes routed by country and period turns a compliance headache into a quantifiable map of where risky cargo concentrates.
• Cross-country benchmarking - Germany at a third of EU performance, Poland second, France third: any pan-European claim about rail competitiveness resolves against one consistent series instead of twenty-seven national publications that each define a tonne differently.
• Investment and macro due diligence - the 2018-to-now arc from 410 billion tkm down to 375 is exactly the kind of structural context a credit or equity thesis needs before it touches a single operator.
• Journalism and academia - official European statistics with the methodology documented, which is as defensible as continental freight attribution gets.

Data scientists and analysts get a clean country-by-period-by-commodity panel whose dimensions repeat across every table in the family - one parser, many datasets, no bespoke cleaning per release. Market researchers and consultants hold the strongest claim of all: an official country-year grid is the benchmark layer every European freight engagement ends up needing, whatever the client sells. Journalists and academics cite the EU's own statistical office when writing about modal share, decarbonisation or industrial geography, which keeps editors and peer reviewers off your back. Developers and builders get typed enums and stable codes that flatten into schemas without scraping worksheets.

It is the authoritative measurement layer for European rail freight - aggregates you can cite, not feeds you can query live. The natural cross-Channel counterpart is the ORR Data Portal - UK Rail Statistics, which goes deeper on its own island with operator-level GB journeys, revenue and punctuality; we break the trade-offs down in the head-to-head comparison. If you need a single global indicator instead of a European grid, the World Bank Railways Goods Transported Indicator trades detail for reach. For the regulatory and safety frame around the network, see the European Union Agency for Railways; for breadth across every transport mode, the EU Open Data Portal - Railway Datasets is the wide-but-shallow catalogue. Teams usually pair this record with one live operational layer: official counts for the trend, feeds for the present.

Four honest notes from our verification pass.

Aggregates all the way down. There are no waybill-level, wagon-level or operator-level records here - the finest grain is country-period-commodity. If your question needs shipment events, this dataset anchors the trend and a different record supplies the events; we will say so up front rather than oversell.

Scope follows the reporting rules. Figures cover main undertakings above regulatory reporting thresholds, so very small operators do not appear, individual cells can be withheld for confidentiality, and Belgium drops out of certain freight tables entirely. Treat thin cells as signals, not noise.

Some cells arrive pending. German dangerous-goods figures were still outstanding at publication, and the quarterly series covers a defined window - Q1 2023 through Q4 2024 - rather than running as deep as the annual tables. If either is your whole question, ask us for current status before you build on it.

Companion families vary in completeness. The goods tables are the solid core; infrastructure and rolling-stock companions were not fully confirmable at research time. We confirm what exists before promising it.

Because "European rail freight statistics" is really a family of tables wearing one badge - units that flip between thousand tonnes and million tonne-kilometres, coverage codes that distinguish inbound, outbound and transit hauls, commodity classifications with their own version history, and confidentiality holes that differ by country. Datadory normalizes the family to the shared dictionary above, keeps successive editions accumulating so revisions become observable history instead of overwritten files, and cuts samples to named countries, commodity groups and coverage splits before any recurring delivery is configured. Browse the rest of the vertical on the rail transportation data hub or the best rail transportation datasets, then request the sample - it ships first either way.