Global Rail Platform

← All country boards · Delay analytics · Live departures · Signalling

Train performance · Japan

Japan train performance.

Live national railway punctuality from ODPT open data. Data updates continuously; this board refreshes every 10 minutes via server cron.

Live rolling window — source: Official live departures feed for Japan, polled server-side every 10 minutes. Not an official published punctuality statistic; see Provenance below.

Live board Next refresh in 10:00 Updated 03:40

Reading the board

Four numbers summarise the rolling window. The same metric is then broken down by delay bucket, by route and by station.

99.2%
Departures on time within 60 seconds
On time≤60s
117
Services running in the current window
Realtimenow
118
Departed stops sampled across 0 hubs
Samplerolling
2.0s
Mean delay per departure
On timeavg

Delay Buckets

How late, and how often

Each departure falls into one bucket by its delay. The bar shows the share of departures in the current window.

On Time (≤60s)117
Minor (1–5m)1
Moderate (5–10m)0
Major (>10m)0

Bucket thresholds are the board\'s own convention: on time ≤ 60s, minor 1–5 minutes, moderate 5–10 minutes, major beyond 10 minutes. Cancelled services are tracked separately and are not folded into these buckets.

Trend

The window and the days behind it

The live window alongside the day-by-day record this board has published since it went live.

Current window

Japan · rolling window
Live
On time (≤60s)99.2%
Major (>10m)0.0%

Rolling window · compiled from the board\'s own endpoint

Day by day

On time % · last 7 recorded days
7-day 85.9%
Sat 14 Nov100.0%
Sun 15 Nov85.9%
Mon 16 Nov78.8%
Tue 17 Nov76.5%
Wed 18 Nov93.0%
Thu 19 Nov83.4%
Fri 20 Nov83.8%
7-day on time
85.9%
30-day on time
90.5%
Days recorded
56

Day-by-day record · one sample per day (the day\'s busiest pull) · not an official statistic.

By Route

Where trains lose time

Service classes ranked by the share of departures on time. Each row carries the sampled stop count and the average and worst delay.

Oedo30 stops sampled
100.0%
avg 2s · max 60s
Blue19 stops sampled
100.0%
avg 0s · max 0s
Mita17 stops sampled
100.0%
avg 0s · max 0s
Arakawa17 stops sampled
100.0%
avg 0s · max 0s
Asakusa15 stops sampled
93.3%
avg 8s · max 120s
Shinjuku15 stops sampled
100.0%
avg 4s · max 60s
Green5 stops sampled
100.0%
avg 0s · max 0s

Top Stations by Volume

The busiest hubs

The same reading, station by station, ordered by how many departures each contributed to the window.

Unknown station118 stops sampled
99.2%
avg 2s · max 120s

Delay Repay

Was your train delayed?

If a journey on this network ran late, you may be owed money back. Check your refund in a couple of minutes — no claim knowledge needed.

Check your refund instantly ↗

Independent helper. Refunds depend on the operator's Delay Repay scheme.

Provenance

About this data

Every board states where its numbers come from and what they mean, in the one place a reader will look for it.

Source
Official live departures feed for Japan, polled server-side every 10 minutes.
Metric
Per-departure delay in seconds (realtime − planned time).
On time
Delay ≤ 60 seconds. Early trains count as on time.
Late
Delay > 60 seconds · cancelled services tracked separately.
Coverage
The monitored hub stations, rolling ~20-minute window (staggered pulls).
Not an official Japan performance statistic. This board publishes the rolling live window only; extended 7- and 30-day views shown in prototypes are labelled samples and are not data.

Another country, the same reading.

Every board reads exactly like this one — one metric, stated plainly, with its own provenance attached.

Browse every board

Help keep this directory complete.

Know a line, place or listing we’re missing — or spotted a mistake? Add it in a minute.

ESC
Go to
↑↓ to move ↵ to open Esc to close
Advertise Sponsorship

Report a problem

Tell us what is not working or looks broken. We attach the page and your browser automatically — no need to describe the technical side.

Routed to the right desk.

Mastodon Lemmy Pixelfed