Estimate a Launch Impact with Difference in Differences
Company: Pinterest
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Onsite
Overview: Estimate a regional product launch effect with difference in differences when retrospective randomization is unavailable. Define the estimand, test parallel-trend credibility, use event-study and placebo diagnostics, cluster uncertainty correctly, and handle spillovers, selection, and staggered adoption.