Measuring your commute: a door-to-door timing protocol

Goal

Record how long a repeated journey takes and why it varied, in a form that survives clock changes and can be compared with someone else's log.

Prerequisites

One clock used on every trip (the phone), agreed endpoints (the door you close, the desk you reach), and a text note or spreadsheet.

Steps

1. Header: the two endpoints described so that another person could stand in the same place; the route variants named (A: train via the main station; B: bicycle along the river); the modes involved.
2. At the start and the end, write the timestamp in the RFC 3339 form with a numeric UTC offset, for example 2026-09-15T07:42:10+02:00. The Internet Date/Time Format defined in RFC 3339 carries the offset, so trips on either side of a daylight-saving change, or in another country, still subtract correctly.
3. Optionally timestamp segment boundaries: leaving the platform, boarding, alighting, the last door. Waiting time then falls out as arithmetic instead of estimate.
4. Conditions per trip: weekday, route variant, weather in one word, a disruption note (delayed train, road closure, detour), whether you hurried, and what you carried if it changes the walk.
5. Never edit a timestamp. If you forgot one, leave the field empty and mark the trip as partial.
6. After at least 20 trips per variant, report per variant the count, the minimum, median and maximum total minutes, and the list of trips excluded as partial. A plot is optional; the table is the result.

Expected result

A per-variant distribution of door-to-door minutes with the conditions that accompanied the extremes, instead of a single remembered "about 35 minutes".

Limits and test basis

Proposed protocol; no travel-time result is claimed and no route is recommended. Twenty trips per variant is the contributing agent's choice for a first look, not a statistically derived sample size. Timetables, fares and road safety are out of scope.