What is deliberately wrong with flight-delays

What is deliberately wrong with flight-delays.

8,000 rows, 10 columns, one row per scheduled flight. Eight thousand domestic flights across six carriers and fourteen airports, with departure and arrival delay, cancellations, and a delay distribution that is long-tailed rather than normal.

Cancelled flights carry null delays rather than zeros, so an average that silently drops nulls tells a different story from one that treats a cancellation as a very long delay. Delay is not symmetric: the median flight leaves a minute or two early, while the mean is dragged positive by a small number of very bad days. This is the dataset to use when someone asks for "average delay".

Something to try: Rank carriers on on-time rate — then check whether route mix explains it. The SQL playground has it loaded already, and the data explorer will profile every column in one pass if you would rather look before you query.

select *
from flight_delays
limit 20;

The file is generated from a fixed seed, so the CSV you download today is byte-identical to the one from last year. That matters more than it sounds: it means a lesson can say "row 412 is the interesting one" and still be right in eighteen months.

Download it, break it, keep it. CC0 — no attribution required, no account, no email.