Skip to content

Method v1.0 · versioned, with a change log

How we measure, in full.

Every report, verification page and corridor page links here and names the version it used. If you want to check one of our numbers, this is where to start.

1. What we measure

Door to door trip time from the home's selected gate to the destination's entrance, split by leg: gate to main road, the main stretch, parking or drop off, walks to and from stations, and waiting time for metro, rail or bus.

2. Sources

  • Road legs: a traffic aware routing source, named on reports once its licence is signed.
  • Metro, suburban rail and bus legs: published GTFS or timetable data where operators publish it under a usable licence.
  • Walking legs: OpenStreetMap paths and station entrances (credited under the ODbL), corrected by field checks.
  • Field Verified trips: surveyor GPS logs with timestamps at each leg.
  • Infrastructure status: the implementing agency's own published status, with the link and the date checked.

See data sources for what we store from each and its known limits.

3. Sampling

Paid reports sample a departure every 15 minutes across each of your windows on 10 working days, using live queries at the actual departure time. The free Commute Check uses the routing source's traffic aware estimates for weekday 8 to 10 am and 6 to 8 pm, shown as a range.

Public holidays, bandhs and declared disruptions are excluded and listed on the report. If a sample fails, it is retried within the same 15 minute slot; if it still fails, it is marked missing, and a trip with more than one missing sample in five on any day says so.

Ten working days × five departure times

Day8:158:308:459:009:15
Mon4348525751
Tue4146556053
Wed4447515649
Thu4249586857
Fri4045495450
Mon4652617263
Tue4247535852
Wed4348505551
Thu4146525954
Fri3944485347

Typical (median): 50.5 min

Bad day (90th percentile): 59 min

Worked example: 50 sampled departures. Typical is the middle value, bad day is the value nine in ten came in under.

4. What the numbers mean

Typical is the median of all samples in the window: half were quicker, half slower. Bad day is the 90th percentile, computed by the nearest rank method: one trip in ten took longer.

We do not report an average. One very long trip pulls an average up, and one empty road pulls it down; neither describes an ordinary day or the day that makes you late.

5. Assumptions

  • Walking speed: 4.5 km per hour unless field measured.
  • Parking or drop off: a stated allowance per destination, or field measured.
  • Metro, rail and bus waiting time: half the published headway, unless a timetable gives exact departures.
  • Lift and security time at large office campuses: included only where field measured.

6. When sources disagree, and known limits

When routing estimates and field trips disagree, both are shown. The report says which one the summary uses and why. Usually it is the field trip, because it starts at the real gate.

Known limits: monsoon, festivals, school vacations, road works started after the measurement, and service changes. A report describes the measured period only. That is why every paid report includes one free remeasurement within 90 days.

7. Independence and corrections

Buyers pay for their reports. Developers pay for measurement under Verified Commute and cannot edit, hide or reorder a measured number; see the verification policy. Nobody can pay for a result.

To report an error, write to support@imperibuild.com with the report reference. Every correction is logged with its date, and the report shows its version history.

8. Change log

VersionDateChange
v1.015 Nov 2026First published version: door to door legs, 15 minute sampling on 10 working days, median and 90th percentile, assumptions as listed above.

Measure the trip you’ll actually make.

Your home, your office, your hours, across ten working days. Typical and bad day time, with the method on the page.

Get a Commute Report for ₹499Or run a free Commute Check