ArrivalQ

Which flight is worth waiting for? Our own experiment.

Tech Prototype · parked Client work · shown re-skinned

1–10

score on every inbound flight

data sources, one board

20 min

the windows that matter

The problem

Airport rideshare queues are a casino — drivers stage for an hour on vibes, and the difference between a good and bad night is which twenty minutes you showed up for.

The data to answer it exists — arrivals, delays, red-eyes, concerts letting out, weather turning — it just lives in five apps that don't talk.

What we built

A scoring engine that rates every inbound flight on time-of-day, status, cluster density, origin type, and event and weather bonuses.
A multi-source pipeline pulling arrivals, events, and weather into one ranked board.
Best-window math: not just 'good flight', but 'stage at 9:40, not 9:00.'
Tests on both halves of the stack, because a scoring engine you can't trust is a horoscope.

FastAPIScoring algorithmMulti-source scrapingNext.jspytest + vitest

In the product

The board

ArrivalQ

BoardWindowsTerminals

9:40pBest window opens

17Flights scored ≥7

+2Event bonus · arena show

9.1Red-eye cluster — three wide-bodies in 22 minT2

7.4Delayed intl — bags mean longer dwellT1

3.2Commuter hop — skip itT2

Terminal-flavored monochrome, one amber signal color — a flight board that thinks in payouts.

The night, plotted

ArrivalQ

TonightHistory

2Prime windows tonight

38Inbounds tracked

RainSurge multiplier likely

Arrivals density across the evening — the windows fall out of the curve.