Flight seat availability data from public booking pages

Flight number, date and cabin

The key for one availability series.

Availability state

Open, limited, sold out, not operated, not on sale. A small, explicit set of states.

Seats-left counter and its wording

Kept with the exact text, so a counter at a price is not read as seats on the aircraft.

Lowest bookable fare and booking class

A step to a higher class is an inventory signal even when the flight never sells out.

Party size searched

Availability depends on it; it belongs in the key.

Days to departure

Sell-out curves are read against lead time.

Calendar vs live source

Marks which observations were confirmed on the live results page.

Observation time

Lets you date the transition from open to sold out.

1. 01

   Load the schedule first

   Build the list of operated flights per route and date from schedule pages. Only those flight-dates get an availability series, which removes most false 'sold out' rows.

2. 02

   Sweep calendars cheaply

   Fetch month or week views per route and party size on rotating exits. Diff each sweep against the last one and queue the flight-dates whose state or lowest fare changed.

3. 03

   Confirm changes on the live results page

   For each queued flight-date, run the logged-out search once on a short sticky session in the route's main selling market and record the state per cabin. Stop at the results or fare-detail page.

   Record a state, not a price19 linescopy

   from enum import Enumclass State(str, Enum):    OPEN = "open"    LIMITED = "limited"        # counter shown, e.g. "3 left at this price"    SOLD_OUT = "sold_out"    NOT_OPERATED = "not_operated"def classify(row) -> State:    if row is None:        return State.NOT_OPERATED    if row.sold_out_label:        return State.SOLD_OUT    if row.seats_left_text:        return State.LIMITED    return State.OPEN# store(flight, date, cabin, party_size, classify(row), row.seats_left_text,#       row.lowest_fare, row.booking_class, source="live", seen_at=now())

4. 04

   Build sell-out curves

   For each flight, find the first observation in each state and express it as days to departure. Aggregated by route and weekday, these curves are what capacity and demand analysts actually use.

5. 05

   Validate against official traffic data

   When load factor statistics for the period are published, compare them with your sell-out rates by route and month. The comparison tells you how far your public signal can be trusted for nowcasting.

Prices are list prices in EUR, excl. VAT. Volume discounts apply automatically at checkout — full ladder on the pricing page.