TrainRouter Atlas — the world's legendary train routes, as open data

961 train routes · 118 countries · ≈ 432,800 km of route — every route with its key facts and hand-traced geometry, from trainrouter.com, the interactive world railway map.

High-speed spines (Eurostar, TGV, Shinkansen, AVE), classic long-distance runs (Trans-Siberian, California Zephyr), night trains (Nightjet, the Ghan) and the scenic lines people fly in just to ride (Glacier Express, Bernina Express, the Jacobite).

🗺️ Explore it interactively: trainrouter.com 🤖 Use it from an AI assistant: free MCP server at https://trainrouter.com/mcp (how to connect) — or run it from this repo: npm install && npm start (mcp/)

Files

Related MCP server: db-mcp

Schema

Quick start

import pandas as pd
routes = pd.read_csv("data/routes.csv")
routes.nsmallest(10, "fame_rank")[["name", "from", "to", "distance_km"]]

import geopandas as gpd
gdf = gpd.read_file("data/routes.geojson")
gdf.plot(column="category", figsize=(16, 8))

Sources, accuracy & provenance

• No upstream dataset. This atlas is hand-built — not derived from OpenStreetMap, Overpass, Wikidata, Natural Earth or any other licensed dataset — so it carries no third-party licensing obligations and the CC BY 4.0 licence is the author's to grant.

• Geometry is hand-traced at map scale: each route's LineString is a set of hand-placed [lon, lat] waypoints following the line's real corridor through its stops — deliberately schematic (smoothed for display), not survey-grade alignment, and not extracted from OSM or any geodata source.

• Facts are approximate published values: distance, top speed, journey time, operator, opening year and approximate ridership are compiled from public references (operators' own figures, timetables, press material, encyclopaedic articles). Individual facts like these aren't copyrightable; the licence covers the curated compilation. Good for exploration and visualization, not operations.

• Curated, not exhaustive: the world's notable routes, not every line on earth. Route names and the one-line highlights are the author's own wording.

• Deliberately excluded (they live only on the site, to keep provenance clean): per-route stories and sights, photos (Wikimedia Commons, separately licensed), and the city-to-city journey guides (derived from public schedule feeds).

• Basemap ≠ data: the interactive site renders these routes over OpenFreeMap (OpenStreetMap) tiles, but those tiles are not part of this dataset — the geometry here is the author's own hand-traced lines. OSM © OpenStreetMap contributors.

License & attribution

• Data (data/): CC BY 4.0 — free to use, share and adapt, with attribution to TrainRouter. A link to https://trainrouter.com (or the specific route page in url) satisfies attribution. (full notice)

• Code (the MCP server in mcp/): MIT.

Also available on

• Kaggle: kaggle.com/datasets/albanius/world-train-routes-trainrouter-atlas

• Hugging Face: huggingface.co/datasets/Flightmussy/trainrouter-atlas

• Zenodo (archived, DOI): doi.org/10.5281/zenodo.21322030 — always resolves to the latest version

Citing

TrainRouter Atlas: the world's legendary train routes (2026). trainrouter.com. DOI: 10.5281/zenodo.21322030. https://github.com/Flightmussy/trainrouter-atlas

Updating

The data is generated from the TrainRouter atlas source. New versions land here first; publishing a GitHub release mints a fresh Zenodo DOI and (once the repo's KAGGLE_API_TOKEN/HF_TOKEN secrets are configured) syncs the Kaggle and Hugging Face mirrors automatically via sync-mirrors.yml.

Also in this repo

• mcp/ — source of the TrainRouter MCP server (live at https://trainrouter.com/mcp, listed in the Official MCP Registry as com.trainrouter/atlas), which serves this atlas as tools for Claude and other MCP clients.