On time = arrived at the scheduled time, up to 3 min late (0 ≤ delay ≤ 3 min).
Early = before scheduled time (delay < 0).
Late = more than 3 min after schedule (delay > 3 min).
Same numerator (arrived at-or-after schedule), but counting anyone who arrived within the larger window.
Within 5 min = 0 ≤ delay ≤ 5 min.
Within 7 min = 0 ≤ delay ≤ 7 min.
Each bar = one whole-minute bucket of the schedule-vs-actual delay, counted across all observed stop arrivals.
Outliers beyond −15 min or +45 min are clamped to the boundary buckets.
On Time Service Delivered = percent of scheduled passenger stop-arrivals that occurred within the rider-experience window: no more than 1 minute early and no more than 7 minutes late (−60 s ≤ delay ≤ 420 s). Route origins, final terminals, and stops where GTFS prohibits pickup are excluded because this tracker cannot measure an arrival there for a waiting rider. Every other scheduled stop counts, including stops on trips that were never observed. Colour scale: green ≥ 95%, orange at 90%, red at 80–75%, darkest red ≤ 65%.
Observed running = trips with at least one recorded stop arrival (regardless of delay). Dropped trips = scheduled trips with no observed arrivals at all.
Trips truncated = trips that ran but were not observed reaching their last passenger-boarding stop before the final terminal. The unreliable final-terminal ping is not used.
For each truncated trip, what fraction of the route to its last boarding stop was observed. A heavy tail near 90–100% means tracking ended near the route terminus; a flatter shape means the trip stopped being observed earlier.
Service regularity
Headway CV means coefficient of variation (standard deviation ÷ mean); lower values mean buses were more evenly spaced.
CV is the coefficient of variation (standard deviation ÷ mean) for each Pacific hour. Lower is more even.
Each bar = total stop arrivals we recorded in a 15-minute window, from PT midnight to midnight, stacked by route. Top-10 routes (by total arrivals for the day) are colored individually; everything else is grouped into Other. Limited non-Transbay routes are excluded.
A "data point" here is one finalized (trip, stop) observation — the same denominator used by every other chart on this page.
For each observed stop arrival, we compare the bus's lateness/earliness to the scheduled gap to its neighboring scheduled trip at the same stop on the same route.
Late buses use the prior scheduled trip's gap; early buses use the next.
Example (late): a 10-min headway with a bus 6 min late = +60% distortion (the gap from the previous bus stretched from 10 to 16 min).
Example (early): a 5-min-early arrival on a 5-min headway = −100% distortion — the bus collapsed onto the previous scheduled slot.
Higher absolute values = more disruptive to riders waiting at the stop.
Click a column header to sort. Click any row to expand it for the full per-route metric list. Headway CV is the coefficient of variation (standard deviation ÷ mean).
Limited routes have fewer than 10 scheduled runs that day. Lettered Transbay routes are always included.
| Route | Δ distribution | On Time Service Delivered % | On time % (≤3 min) | p50 delay (min) | Headway CV | Bunching status |
|---|
Data pipeline
Request success and latency are reported in the weekly view so short collection windows are not mistaken for missing service data.