GTFS (Parquet)
SNCB: belgium's national train operator
- Format
- GTFS Parquet
- Updates
- Daily (regenerated from the GTFS feed)
- Latest snapshot
- Snapshots stored
- about 693
curl -H "Authorization: Bearer $MOBILITYTWIN_TOKEN" -o gtfs-parquet.zip "https://api.mobilitytwin.brussels/sncb/gtfs-parquet"The SNCB/NMBS timetable with one Parquet file per GTFS table, in a zip, 40-75% smaller than the GTFS zip. Polars or DuckDB can read single columns without loading a whole table. Built with gtfs-parquet v0.4.0. No shapes table: SNCB publishes no shapes.txt.
A zip with one Parquet file per GTFS table.
Which days have data
Data from 21 Aug 2024 to 2 Nov 2026, on 790 days. 30 of the last 30 days have data. A typical day holds 1 snapshot. Days are in UTC.
2026
JanFebMarAprMayJunJulAugSepOctNovDec
2025
JanFebMarAprMayJunJulAugSepOctNovDec
2024
JanFebMarAprMayJunJulAugSepOctNovDec
- Full day
- Partial, 50–90% of a typical day
- Low, under 50%
- No snapshot
- Not expected
Click a day above, or pick a date, to get its requests.
Parameters
| Name | Meaning |
|---|---|
timestamp | Return the last snapshot at or before this moment, in Unix seconds or ISO 8601 (e.g. 2026-05-06T12:00:00Z). Without it, the latest snapshot is returned. |
Python example
# pip install gtfs-parquet>=0.4.0
import requests
import tempfile
from gtfs_parquet import read_parquet
from gtfs_parquet.ops.network import describe
from gtfs_parquet.ops.calendar import get_first_week, compute_busiest_date
from gtfs_parquet.ops.routes import compute_route_stats
from datetime import datetime, timedelta, timezone
yesterday_noon = (datetime.now(timezone.utc) - timedelta(days=1)).replace(
hour=12, minute=0, second=0, microsecond=0
)
params = {
"timestamp": yesterday_noon.isoformat(), # ISO accepted; epoch also works
}
url = "https://api.mobilitytwin.brussels/sncb/gtfs-parquet"
data = requests.get(url, headers={
'Authorization': 'Bearer [MY_API_KEY]'
}, params=params).content
with tempfile.NamedTemporaryFile(delete=False, suffix=".zip") as f:
f.write(data)
feed = read_parquet(f.name)
# Quick summary of the feed
print(describe(feed))
# Find the busiest date in the first week
week = get_first_week(feed)
busiest = compute_busiest_date(feed, week)
print(f"Busiest date: {busiest}")
# Compute per-route statistics for that week
route_stats = compute_route_stats(feed, week)
print(route_stats.sort("num_trips", descending=True))