S06 · interview simulation

Python gym: turning cumulative booking curves into pickup in pandas

People, data and outputs in this scenario are simulated. Each stage is recorded when you commit; the reveal comes after.

5stages
35 minabout
0stages committed

The brief

A Python interview (simulated) with the course's synthetic booking_curve joined to route and market. In 2024 it has 90,402 rows: one row per flight and departure date and 13 snapshots at 90, 60, 45, 30, 21, 14, 10, 7, 5, 3, 2, 1 and 0 days before departure.

Columns: date, flight_no, route, market, days_to_departure, bookings (bookings made since the previous snapshot) and cumulative_bookings. The frame is passed in shuffled; no row order may be assumed.

Example, flight ZX104 departing 2024-07-15:

days_to_departure	90	60	45	30	21	14	10	7	5	3	2	1	0
bookings	3	14	22	27	23	22	12	9	9	7	7	4	2
cumulative_bookings	3	17	39	66	89	111	123	132	141	148	155	159	161

The interviewer: "What share of final bookings has a flight typically made by 30 days before departure?"

Which implementation answers it, per flight and departure date?

Bookings in the 30-day snapshot over bookings in the final snapshot: `bc[bc.days_to_departure == 30].bookings.sum() / bc[bc.days_to_departure == 0].bookings.sum()`Pivot cumulative bookings by snapshot and divide the 30-day column by the 0-day column: `p = bc.pivot(index=["flight_no", "date"], columns="days_to_departure", values="cumulative_bookings")`, then `share = p[30] / p[0]`.The share of all bookings made inside the last 30 days of the curve, from the flow column, which needs no pivot and reads naturally as the last month of sales: `bc[bc.days_to_departure <= 30].bookings.sum() / bc.bookings.sum()`The fourth snapshot of each group over its maximum, which avoids a pivot: `bc.groupby(["flight_no", "date"]).cumulative_bookings.nth(3) / bc.groupby(["flight_no", "date"]).cumulative_bookings.max()`

The brief

Stage 1 of 5