A stop’s record is counted from the buses that called there in the last 30 days.

What we count

• Every time a bus was due at this stop in the last 30 days, we judged it against the timetable. On time means from one minute early to five minutes late.
• A trip that never came counts as a miss. A call we could not judge is left out. Only days we captured in full are counted, and the page says how many.

The hour chart and the weak window

• The chart shows weekdays only. An hour with too few judged calls shows no bar and says it is not measured yet.
• The weakest window is the lowest short run of hours in a row, with enough judged calls between them. Worst and best route each need enough judged calls.

The main cause

• The main cause is explained, not measured. A rule-based classifier looks for patterns in this stop’s own counts and names a cause only when its tests pass with at least medium confidence.
• When the evidence does not support a cause, the page says so. We never name one by guessing.

What is left out

• Only timetable timepoints are stored. A stop that is not one has no record, and the page says it is not measured yet. A stop with too few judged calls gets no share and no grade.

Where the numbers come from

• Our stop visits, judged against the timetable

Method version and change log

A published month never changes. If a method is revised, past months keep the method they were worked out with.

• Version 1 · October 29, 2026. First published. A stop’s own on-time record for 30 days, by hour, with its grade, its worst and best route and its main cause.