120 refund requests to the homeware shop of tickets: what the customer
wrote, what the order system knows, the item's state, and the decision.
The same table as Python's functai.datasets.refunds().
Format
A tibble with 120 rows: id, message (what the customer wrote),
item, price (dollars), days_since_delivery (from the order system),
final_sale (logical), state (a factor: unopened, opened_unused,
used, damaged, wrong_item, faulty) and decision (a factor:
approve, deny).
Refund rules
decision follows them exactly:
Damaged on arrival, or the wrong item (or part of the order missing): a refund within 60 days of delivery, final sale or not.
Faulty (it failed in normal use): a refund within 365 days, final sale or not.
Unopened, or opened but not used, and no longer wanted: a refund within 30 days, and never for a final-sale item.
Used and no longer wanted: no refund.
The messages were written by a language model from each row's facts, in
varied tones and lengths; some say what happened only indirectly. The
facts, and so the decisions, were drawn first (data-raw/refunds.R).