Public methodology register
Buses arriving bunched
- Version
- settled-reliability-patterns-daily-v1
- ID
- route-regularity:bunched-gap-share
Definition
- Definition
- The share of back-to-back buses on a route that arrived much closer together than the timetable planned for those same two trips.
- Formula
- At each stop, buses of one route and direction are put in the order they arrived. For each pair, the observed gap is compared with the scheduled gap between the same two trips. Bunched: observed gap under 0.25 times the scheduled gap. Long gap: over 1.75 times. Share = bunched pairs divided by judged pairs.
- Units
- percent of judged pairs, with the count of pairs
- Population
- Pairs of buses on the same route and direction that arrived one after the other at the same stop, from settled final predictions.
- Measurement window
- The 28 most recent completed service days, by calendar date in Ottawa.
- Coverage
- Only days the daily rollup computed. Days with partial feed capture and days when the live trip numbers did not match the timetable are left out and listed.
- Publication eligibility
- A pair is judged only when the timetable planned 2 to 30 minutes between those two trips. A route needs at least 30 judged pairs before a share is shown.
- Scope
- Every bus route with judged pairs, by hour band: early 4 to 6, morning rush 6 to 9, midday 9 to 15, afternoon rush 15 to 18, evening 18 to 22, night 22 to 4 (Ottawa time).
- What this number can support
- A high share means buses often come together and then leave a longer wait behind them. It is not a count of late buses.
Implementation status
- Source candidate
- Implemented in this source candidate.
- Public production
- This metric version has not been independently verified in the public production deployment.
The daily calculation is implemented in this source candidate (worker rollup, migration 0396). Public production verification for this version is not yet recorded.
Sources and identifiers
- Method
settled-reliability-patterns-daily-v1Daily reliability-pattern method - Current API
reliability-patternsBunching, slow segments and on-time by hour
Exclusions
- OC Transpo O-Train lines are left out: the operator publishes no live train positions, so their arrivals repeat the timetable.
- Comparing each pair with its own planned gap means an unseen bus in between, or a bus passing another, does not create a false bunch.
Uncertainty and interpretation
- Arrival times are the last prediction made no earlier than 10 minutes before the scheduled time, not a recorded door opening.
Current evidence
No registered drill-down record is available for this metric. The public page remains the current presentation source.
Change history
No registered material change is recorded for this metric version.