Diagnose 4% weekly revenue drop using history
Company: Instacart
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Technical Screen
Overview: Diagnose a 4% week-over-week revenue drop using only 104 weeks of weekly totals, deciding whether it is ordinary noise or a genuine anomaly and how unusual it is. Handle seasonality, moving holidays, and a single prior observation of the same week, then extend the analysis with geo-level orders, order value, users, and marketing spend.
Read the full Instacart Data Scientist interview experience this question came from