The aircraft before your flight / July 2025
Your planea day. One aircraft left San Francisco two and a half hours late. Its next departure was only eleven minutes late. Follow the plane, and a delay becomes a sequence.
SFO → HNL +150min
HNL → PPG +11min
PPG → HNL +12min
Tail N393HA · flights scheduled July 10, 2025 · minutes relative to each departure’s own schedule
Follow its day ↓ 01 / A DELAY IS NOT A DESTINY
There is another timetable behind your boarding pass. The plane at your gate may already have made several trips. But tracing those trips does not mean adding every late minute together. Each flight has its own scheduled departure, its own arrival, and a planned interval on the ground. The useful question is how the difference from that schedule changes as the aircraft moves.
Here are three real sequences reported under Hawaiian’s HA carrier code to the Bureau of Transportation Statistics. They come from a frozen July 2025 extract, not a live flight tracker. The journeys are deliberately short enough to follow: within each pattern, we selected the eligible day with the fewest legs, then the earliest date and tail identifier to break ties.
01 Near schedule Los Angeles to Maui, then Seattle and Anchorage: departures 4 minutes early, 5 early, and 9 late. At the last arrival, the aircraft was 13 minutes early.
02 Delay, then recovery San Francisco to Honolulu, then Pago Pago and back: 150 minutes late became 11, then 12. The last arrival was on July 11, two minutes early.
03 Delay that persists Honolulu to San Diego and back, then another San Diego flight: 2 minutes early, 62 late, then 35 late. Some delay was absorbed; a late departure remained.
Recovery and persistence can coexist. In the third story, the last departure was 27 minutes less late than the one before it. It still left 35 minutes behind schedule. A useful description keeps both facts.
02 / FOLLOW THE SAME AIRCRAFT
The day, leg by leg. N393HA, 2025-07-10: 3 reported flights. Departure deviations: +150, +11, +12 minutes. Delay, then recovery.
View Clock labels
Outline: scheduled flight interval. Solid: actual interval. The ribbon uses UTC; the itinerary below offers local clocks. On a phone, follow the vertical itinerary.
Loading the chart enhancement. All times are in the itinerary below.
Before this window:HA 12 · HNL → SFO, scheduled date 2025-07-09. Tail, airport and chronology are consistent with the first leg shown.
Scheduled departure Jul 10, 09:35 PDT
Actual departure Jul 10, 12:05 PDT
Arrivals and original reporting fields Scheduled arrival: Jul 10, 12:00 HST
Signed arrival deviation: +154 min. Scheduled / actual gate-to-gate durations:325 /329 min.
Reported categories on this row: CarrierDelay 154 min; WeatherDelay 0 min; NASDelay 0 min; SecurityDelay 0 min; LateAircraftDelay 0 min. These are reporting categories, not a reconstruction of the aircraft's whole day.
Cancellation flag: 0.00; diversion flag: 0.00. BTS archive CSV record line 584013.
2025-07-10~HA~11~14771~12173~0935127 min on the ground at HNL between the reported flights
Scheduled departure Jul 10, 16:30 HST
Actual departure Jul 10, 16:41 HST
Arrivals and original reporting fields Scheduled arrival: Jul 10, 21:10 GMT-11
Signed arrival deviation: -1 min. Scheduled / actual gate-to-gate durations:340 /328 min.
No delay-cause values were supplied on this row. Blank is not evidence of no cause.
Cancellation flag: 0.00; diversion flag: 0.00. BTS archive CSV record line 589260.
2025-07-10~HA~465~12173~14222~1630133 min on the ground at PPG between the reported flights
Scheduled departure Jul 10, 23:10 GMT-11
Actual departure Jul 10, 23:22 GMT-11
Arrivals and original reporting fields Scheduled arrival: Jul 11, 05:40 HST
Signed arrival deviation: -2 min. Scheduled / actual gate-to-gate durations:330 /316 min.
No delay-cause values were supplied on this row. Blank is not evidence of no cause.
Cancellation flag: 0.00; diversion flag: 0.00. BTS archive CSV record line 589274.
2025-07-10~HA~466~14222~12173~2310After this window:HA 628 · HNL → SAN, scheduled date 2025-07-11. Tail, airport and chronology are consistent with the last leg shown.This window does not describe every use of the aircraft.
PINNED FLIGHT / COMPUTED OBSERVATION
HA 465 · HNL → PPG HA 465 · HNL → PPG departed 11 min late. Its signed departure deviation was 139 minutes lower than the preceding departure (+11 − +150 = -139 min). This compares schedules, not causes.
The observation is calculated from signed BTS departure delays. It does not attribute the change to weather, a crew, air traffic control, or the prior flight.
Your local note about this flight 03 / THREE DAYS ARE NOT THE WHOLE MONTH
How often did these patterns appear? Among 660 eligible aircraft-days , 334 stayed near schedule, 11 met the recovery rule, and 24 met the persistence rule. The remaining 291 eligible days followed other patterns. These are descriptions of this selected cohort, not Hawaiian’s official on-time performance rate.
334 / 660 Near schedule ·50.6%
11 / 660 Delay, then recovery ·1.7%
24 / 660 Delay that persists ·3.6%
291 / 660 Other eligible patterns ·44.1%
The denominator matters. The July 2–30 window contains 1,627 aircraft-days with a tail number. We exclude 967: 941 have fewer than three reported flights, 48 have a reporting or continuity break, and 22 are in both groups. Short mainland journeys and unreported activity therefore affect who enters this comparison. The featured three-leg days are not a random sample of the eligible days, many of which have much longer inter-island itineraries.
Comparison pattern All checked, internally continuous sequences of at least three reported flights.660 aircraft-days selected.
Departure distribution is loading.
Read the distribution counts Early (<0):2439checked flight rows 0–14 min:1694checked flight rows 15–59 min:616checked flight rows 60–119 min:70checked flight rows 120+ min:28checked flight rows Each flight contributes once. Missing actual times are excluded; other flights on an ineligible day can still have valid reported times.
Explore an aircraft-day 04 / THE PART THE TIMETABLE CANNOT TELL
A sequence is evidence. A cause needs more. In the recovery story, the first arrival was 154 minutes late. There was a scheduled 270-minute gap before the next departure. The actual gap was 127 minutes. That arithmetic helps explain how the later departure could be much closer to schedule. It does not tell us what people did during that interval or why the first flight was delayed.
BTS does publish carrier-reported delay categories. You can inspect the values under each flight. A blank category remains blank; we do not replace it with a story about weather or a late inbound plane. Even a reported late-aircraft value describes that reporting category for a particular flight. It cannot identify every earlier event that contributed to the day.
A repeated tail number is also a limited kind of evidence. The adapter checks airport continuity and chronology, and stops at unresolved records or a gap longer than twelve hours. Reporting omissions, positioning flights, international activity, and aircraft changes can still be outside the extract. “No break detected” means the available rows fit the checks; it is not proof that nothing else happened.
What can you take away? Delay is a changing distance from a series of schedules. An aircraft can start late and come close to schedule, or recover time and still remain late. Following a reported sequence makes those distinctions visible. It cannot tell you when your next flight will leave.
05 / TAKE THE FLIGHT WITH YOU
Same flight. Your note. Another screen. Pin a flight, attach a note, then reopen its link or import its packet in another browser session. Notes stay attached to the flight’s identity and edition. They remain your unreviewed text, separate from the computed observations. Nothing is submitted to a comment service. The link contains your notes, so share it only with the people you intend to read them.
SOURCE NOTES / SAVED HISTORICAL EDITION
The records behind the ribbon. Historical HA reporting-carrier records, July 2–30, 2025. Observed sequences do not establish causes or predict a departure. This edition has no recorded successor or correction. Retrieved2026-09-05T19:26:56.853Z; edition ha-2025-07-74309a3a734d-v1. The source does not supply per-row publication or revision timestamps. A saved file cannot update itself.
BTS field definitions ·Reporting-carrier table coverage · BTS public access guidance
Method, acceptance evidence, and how this was made Aircraft-day means one tail number and the scheduled origin-local flight date, not midnight-to-midnight UTC. July 1 and 31 are boundary context, outside the July 2–30 comparison. Forty-eight rows in the full month have no tail number. Each composite flight ID uses date, reporting carrier, flight number, origin/destination airport IDs, and scheduled departure. Duplicate IDs stop the build.
Instants use a versioned mapping of 21 airport IDs to IANA zones and timezone database2025c. Scheduled departure plus signed delay determines actual departure; elapsed durations determine arrivals, which must agree with the reported local clocks and arrival delay. Missing, cancelled, diverted and impossible connections are visible breaks.
Pure example adapters supply the itinerary, chart data, sheet and packet. Semiotic’s high-level charts and custom layout components render the ribbon, network and comparison, with Artifact Contracts for the exported evidence. The host owns file/URL storage. These helpers stay example-local until another materially different story establishes a reusable public API.
Automated checks compare the nine featured source rows with an independent time calculation and test flight identity, cross-session notes, chart evidence, keyboard controls and four viewport widths. A reproducible production-route measurement uses a throttled desktop phone viewport; it does not establish performance on a phone.
This is E02’s first implementation. Editorial review, real Android performance, assistive-technology sessions, five-reader acceptance, and the full shared gates remain open. The strategy brief records automated checks and remaining work; this page does not claim those human or device gates have passed.