World Bank Railways Goods Transported Indicator

It is the World Development Indicators series that measures how much freight each economy moves by rail: goods transported, expressed as metric tons multiplied by kilometres travelled and reported in million ton-km under code IS.RRS.GOOD.MT.K6. One annual national value per country-year spans roughly a hundred economies and aggregates from 1995 through 2021, sitting in the Infrastructure topic of the World Bank's flagship statistical collection.

The scale is the story. China reports 3,018,200 million ton-km for 2019, the United States 2,239,400.714 for 2021 and India 719,762 for 2021 - the three heaviest rail freight systems on earth, measured on the one unit that makes them comparable. If you benchmark corridors, size rolling-stock demand or stress-test an industrial-activity claim, these are the figures everyone else quotes. Get a sample of this dataset and we will return rows shaped exactly like the dictionary below.

Every observation hangs off the same three-column spine - economy, year, tonne-kilometres - so once one country parses, the whole series parses. Latest published values for the three largest systems:

# IS.RRS.GOOD.MT.K6 -- Railways, goods transported (million ton-km)
country         countryiso3code   date   value
China           CHN               2019   3,018,200
United States   USA               2021   2,239,400.714
India           IND               2021   719,762

# every row carries its provenance alongside the number
indicator.id: IS.RRS.GOOD.MT.K6
indicator.value: "Railways, goods transported (million ton-km)"
sourceOrganization: "Railisa Database (UIC); OECD Statistics"

# unreported country-years ship as nulls, never as zeros
date: <year>   value: null

Those are the collection's own recorded values, not projections. Slots marked as a shape preview fill in per record once we cut your sample. Because the spine never widens - no subnational, operator or commodity dimension appears at any point - joining this series to any other country-keyed table is a one-key merge on ISO3 code plus year.

Seven fields anchor the dictionary, and each definition below comes straight from what the records actually carry; the example column shows the field populated for the largest reporting system.

Additional fields on request: the companion series that make intensity work possible - rail passengers carried in passenger-km and rail lines total route-km - fold into the same country-year spine when you take the wider rail bundle. Companion columns are confirmed against live records when we prepare your sample, which is also where column-level naming gets locked down for your pipeline.

Three chips summarize the footprint:

• Geography: worldwide - every economy and aggregate reporting rail freight through the International Union of Railways and OECD channels, roughly a hundred-plus countries in a given year, from the largest systems down to single-corridor networks.
• Time frame: 1995 through 2021 as published, one observation window per year. A quarter-century of history means trend work, cycle analysis and pre/post comparisons run without stitching archives together.
• Granularity: annual national totals, one value per country-year, in million ton-km. There is deliberately no subnational, operator or commodity split - this is the cleanest possible cross-country comparison layer.

Set against the wider Datadory catalog - where the average quality score across all 1,744 datasets is 7.81 - this slice scores 7/10, a band shared by 296 datasets, and by just two of the fifteen rail transportation slices we catalog. Verified field definitions carry the score.

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

You choose the channel and the cadence; the field dictionary above travels unchanged across all three. Structured payloads suit live dashboards, bulk files suit overnight warehouse loads, and spreadsheet-, CSV- and SQL-friendly extracts suit analysts who live in a BI tool. Cadence changes are a settings conversation, not a re-integration project - tell us which economies and which years you need and how often, and the feed follows.

Ranked by how directly the series answers their day job:

1. Market researchers & consultants. Country-level tonne-kilometres quantify rail freight markets where operator-level sources do not reach, giving sizing inputs that hold up across borders. See rail transportation data for market researchers.
2. Journalists, academics & students. A citable global series on what each economy moved by rail, back to 1995 - the reference point whenever modal-shift, decarbonization or industrial-policy stories need hard figures. See rail transportation data for journalists and academics.
3. Data scientists & ML engineers. An exogenous feature with genuine signal: rail freight tracks industrial activity closely enough to serve as a physical-economy input to demand models. See rail transportation data for data scientists.
4. Investors & quant researchers. Tonne-km trends read capacity against output, making the series a GDP-adjacent check on growth claims in freight-dependent economies.
5. Developers & builders. A three-column spine keyed on ISO3 code plus year makes scripted ingestion a configuration exercise rather than a project.

For contrast inside the same industry: Eurostat Railway Freight Transport Statistics goes deep on Europe with commodity and vehicle detail, while this series supplies the global baseline - and the ORR Data Portal covers Britain in official depth.

Notes that pair well with this page:

• Rail transportation data hub - the full pooled view of the industry, from North American network geometry to live European departures.
• vs Eurostat Railway Freight Transport Statistics - the global country-year baseline against the European commodity-depth series, on one comparison page.
• ORR Data Portal - UK Rail Statistics - Britain's official rail picture, including freight, at national depth.
• Best rail transportation datasets - where this slice ranks in the industry and why the score band matters.
• Persona pages - what market-researcher, data-science and journalist teams each do with this series.