Advanced Regression Recipes

Predict a continuous value per entity — time-to-event predictions, counts, durations, and peak values.

Common advanced patterns

Time-to-event — compute days until first occurrence of a condition (complaint, new category purchase)
Cross-channel counting — match events across data sources to count specific behaviors (buy online, return in-store)
Duration filtering — filter events by duration thresholds before aggregating
Set-based tracking — track seen items to detect "new" events (lessons, categories)
Pandas groupby — use pandas for daily/weekly aggregations not supported natively

Recipe	Industry	Advanced Pattern
Days Until Complaint	General	Time-to-event, min timestamp
Buy Online Return In-Store	Retail	Cross-channel order matching
New Category Purchase Time	Retail	Set tracking, timestamp iteration
Peak Daily Data Usage	Telecom	Pandas daily groupby
Days with Long Calls	Telecom	Duration filtering, unique day counting
Training Duration	Fitness	Duration filtering, sum aggregation
New Lesson Duration	EdTech	Set tracking, repeat filtering