All stations in Belgium

We try to maintain a list of all the stations in Belgium using CSV so everyone can help to maintain it on github. Furthermore, we have a PHP composer/packagist library for you to go from station name to ID and vice versa and we convert the CSV file to JSON-LD for maximum semantic interoperability.

Fields we collect

stations.csv

This file describes all NMBS/SNCB stations in Belgium. A station can have multiple platforms (stops), which are described in stops.csv.

• URI: this is the URI where we can find more information (such as the real-time departures) about this station (this already contains the ID of the NMBS/SNCB as well)
• longitude: the longitude of the station
• latitude: the latitude of the station
• name: the most neutral name of the station (e.g., in Wallonia use the French name, for Brussels use both, for Flanders use nl name)
• alternative-fr: alt. name in French, if available
• alternative-nl: alt. name in Dutch, if available
• alternative-de: alt. name in German, if available
• alternative-en: alt. name in English, if available
• taf-tap-code: TSI TAF/TAP number for the station Source: Infrabel Operational point of the railway network .
• telegraph-code: abbreviated name for Belgian stations. Since the split of the former NMBS these abbreviations are no longer assigned consistently, and Infrabel and NMBS may use different ones for the same station: Noorderkempen is NDK in Infrabel publications but NDKMP within NMBS. This column follows Infrabel's form (so, NDK), matching the source below. Source: Infrabel Network Statement, attachment D file D06-RIEI-12_N_20230915.pdf, list 12, cross-checked against the symbolicname field of Infrabel's operational points dataset.
• country-code: the code of the country the station belongs to
• avg_stop_times: the average number of vehicles stopping each day in this station (computed field)
• official_transfer_time: the time needed for an average person to make a transfer in this station, according to official sources (NMBS/SNCB) (computed field)

stops.csv

This file describes all NMBS/SNCB stops in Belgium. Each platform is a separate stop location. All fields are computed using gtfs_data_extractor.php.

• URI: this is the URI where we can find more information about this stop/platform (exists out of URI of the parent station + '#' + platform code)
• parent_stop: this is the URI of the parent stop defined in stations.csv
• longitude: the longitude of the stop
• latitude: the latitude of the stop
• name: stop name
• alternative-fr: alt. name in French, if available
• alternative-nl: alt. name in Dutch, if available
• alternative-de: alt. name in German, if available
• alternative-en: alt. name in English, if available
• platform: the platform code (can also consist of letters, so do not treat this as a number!)

facilities.csv

This file describes facilities available in NMBS/SNCB stations. All fields are computed using facilities_extractor.js. Not every field can still be sourced, and not every one is equally reliable: see Where facility data comes from.

• URI: The URI identifying this station.
• name: The name of this station.
• street: The street of this station's address.
• zip: The postal code of this station's address.
• city: The city of this station's address.
• ticket_vending_machine: Whether or not ticket vending machines are available. Note: Ticket vending machines might be located inside a building (and can be locked when the station is closed).
• luggage_lockers: Whether or not luggage lockers are available.
• free_parking: Whether or not free parking spots are available.
• taxi: Whether or not parking spots for taxis / waiting taxis are available.
• bicycle_spots: Whether or not bicycle parking spots are available.
• blue-bike: Whether or not the has blue-bikes (rental bikes).
• bus: Whether or not transferring to a bus line is possible in this station.
• tram: Whether or not transferring to a tram line is possible in this station.
• metro: Whether or not transferring to a metro line is possible in this station.
• wheelchair_available: Whether or not the station has wheelchairs available.
• ramp: Whether or not the station has a ramp for wheelchair users to board a train.
• disabled_parking_spots: The number of reserved parking spots for travellers with a disability.
• platform_height_cm: The nominal platform height(s) in centimetres above the rail, from Infrabel. One of 28 (the historic model, being phased out), 55 or 76. Stations whose platforms are not all the same height list each value, ascending and ; separated, for example 28;76. Empty when Infrabel publishes nothing for the station. These are nominal figures: an individual platform may be built slightly differently. Replaces the old elevated_platform flag.
• escalator_up: Whether or not the station has an ascending escalator from or to the platform(s).
• escalator_down: Whether or not the station has a descending escalator from or to the platform(s).
• elevator_platform: Whether or not the station has an elevator to the platform(s).
• audio_induction_loop : Whether or not an Audio induction loop (Dutch: Ringleiding) is available.
• sales_open_monday - sales_open_sunday: The time at which ticket boots open on this day of the week.
• sales_close_monday -sales_close_sunday:The time at which ticket boots close on this day of the week.

embarkment_statistics.csv

This file describes how many people embark a train at the given station per day. Statistics from october 2023, source: belgiantrain.be, Cijfers opgestapte reizigers

• URI: The URI identifying this station.
• name: The name of this station.
• nmbs_short_name: The short name of this station, as used in the original source data. For verification and future update purposes.
• average_weekday_embarkments: The average number of travelers embarking on a weekday. Empty when a station does not have traffic on weekdays.
• average_saturday_embarkments: The average number of travelers embarking on a saturday. Empty when a station does not have traffic on saturdays.
• average_sunday_embarkments: The average number of travelers embarking on a sunday. Empty when a station does not have traffic on sundays.

How we collect data

This repository contains scripts which can load all data from the NMBS GTFS public data and from several open data sources. These scripts can be used to generate all CSV files from scratch, and to update existing files.

• bin/gtfs_data_extractor.php builds stations.csv and stops.csv from the NMBS GTFS feed.
• bin/facilities_extractor.js builds facilities.csv (see Where facility data comes from).
• bin/verify_facilities.js checks that facilities.csv is still well formed. Run it after the extractor.

Manual changes and corrections can be made to stations.csv. It is recommended to use the stations.csv file in this repository as a starting point instead of using the scripts to generate this file, as the repository versions includes manual fixes to station names and translations.

Any changes made to stops.csv or facilities.csv will be overwritten by the scripts. Therefore, any pull requests with the sole purpose of updating/modifying these files won't be accepted

Missing stations and missing fields in stations.csv are automatically added when the gtfs_data_extractor tool runs.

How to make a correction

Corrections to names, translations and locations can be made by adjusting fields in stations.csv:

• Names or translations will never be overwritten by the scripts.
• Names in facilities.csv or stops.csv are derived from the names in stations.csv, meaning you only need to update stations.csv.
• The GTFS data extractor script will warn on wrong locations, but won't correct them.

If you want to make a correction to facilities.csv or stops.csv, don't fix the files, but fix the scripts instead, and let these scripts run to update the file for you.

For facility data, the fix usually belongs in OpenStreetMap rather than here. If a station is missing its lift or its bicycle parking, map it there and the next run will pick it up, which helps everyone using that data and not just this project.

Where facility data comes from

facilities.csv used to be scraped from the NMBS website, one page per station. That site was retired, and its replacement is behind Cloudflare, which serves a captcha to anything that isn't a browser. The data went stale in June 2023 as a result. There is no single replacement source, so the file is now assembled from several open ones:

Source	Provides
OpenStreetMap via the Overpass API	Most facilities, matched to our stations on uic_ref
Infrabel Open Data	platform_height_cm, from the nominal platform heights (CC0)
Blue-bike API	blue-bike

To refresh it:

node bin/facilities_extractor.js
node bin/verify_facilities.js

A GitHub Action runs both on the first of each month and opens a pull request with the result, so the data cannot quietly rot again.

What this does not cover

Some columns were only ever published by that NMBS website and have no open source today: street, zip, city, wheelchair_available, ramp and audio_induction_loop. The extractor leaves their previous values alone rather than blanking them, so they still date from June 2023.

The sales_open_* and sales_close_* columns are only refreshed for the handful of stations whose NMBS ticket office publishes its hours in OpenStreetMap. NMBS has also closed most staffed counters since 2023, so treat this column with care.

Where OpenStreetMap maps a feature only patchily, a missing feature is not taken to mean the station lacks it: the previous value is kept instead of writing a confident 0. facilities_coverage.md is regenerated on every run and records, per column, where the data came from and how much of it is still carried over.

A note on licensing

Infrabel publishes its open data under CC0, so platform_height_cm carries no restriction. The other two do.

This repository is CC0, but two of the sources above are not: OpenStreetMap is ODbL (share-alike) and the Blue-bike feed is registered as non-commercial on transportdata.be. Re-publishing their data as CC0 is therefore not clearly permitted. This is unresolved and needs a maintainer decision; it is flagged here rather than left implicit.

In case you just want to reuse the data

Latest update over HTTP

JSON-LD is available at https://irail.be/stations/NMBS if you add the right accept header. For example, using curl on the command line, you would do this:

curl -H "accept: application/json" https://irail.be/stations/NMBS

If you want to change this output, please change the CSV files over here first (we love pull requests)

In PHP project

Using composer (mind that we also require nodejs to be installed on your system in order to use the linked-data interface):

composer require irail/stations

Then you can use the stations in your code as follows:

use irail\stations\Stations;
use irail\stations\StationsLd;
// Using the strongly typed CSV parser:
// getStations() returns a irail\stations\Station array
$brusselsnorth = Stations::getStations("Brussels North")[0];
// getStationByID($id) returns a irail\stations\Station object with the station or null
$ghentstpieters = Stations::getStationByID("http://irail.be/stations/NMBS/008892007");

// Using the linked-data interface
// getStations() returns a json-ld document
$brusselsnorth = StationsLd::getStations("Brussels North")->{"@graph"}[0];
// getStationByID($id) returns a simple object with the station or null
$ghentstpieters = StationsLd::getStationByID("http://irail.be/stations/NMBS/008892007");

Don't forget to do a composer update from time to time to update the data

Building the RDF or JSON-LD

Using scripts, this data can be converted to JSON-LD. In order to run the script, run this command:

First time run this in your terminal (nodejs needs to be installed on your system):

npm install

Or install it globally using the npm package (you will need to run this again when there's an update to the stations file):

npm install -g irail-stations

From then on, you can always run:

# using this repo
./bin/build.js
# or with the global package:
irail-stations

For extra commands, check:

./bin/build.js --help
# or
irail-stations --help

We currently support the output formats TRiG, N-Quads and JSON-LD (default)

License

CC0: This dataset belongs to the public domain. You're free to reuse it without any restrictions whatsoever.

If you contribute to this repository, you agree that your contributions will be licensed under the CC0 open data license.

We do appreciate a link back to this repository, or a mention of the iRail project.