From the data analysis course.
Files
- The Data Analyst Round — the SQL screen (window functions, fan-out), the diagnostic case (“revenue dropped 20% — investigate”), metrics and product questions, and the communication question that most separates candidates. 🔥 marks what comes up constantly
The scope note — read this first
INTERVIEW.md states a principle this bank sits against: “an interview bank for a subject you haven’t practised would be memorisation, not preparation.”
That holds, and this bank exists anyway — written from the course, not from having sat these interviews, and honestly labelled. A map of what the round asks, not a substitute for reps.
What actually prepares you is analyses you’ve done — because the diagnostic case (“why did this drop?”) and the SQL screen reward pattern fluency you only get from having explained real changes in real data → data analysis projects. The “full analyst deliverable” project covers most of this round in one go.
What this round tests differently
It’s half live-SQL, half case study — an unusual shape. The SQL half is fast and mechanical (window functions, and the instinct to sanity-check a number against a known source); the case half is about structure — running a diagnostic as a method rather than guessing, and thinking in decisions rather than numbers.
Two failure modes this bank is written against:
- Freezing on live SQL. The window-function questions come up constantly and reward practice, not theory — write them until they’re reflexive
- Answering with a number instead of a decision. “What’s the metric?” and “how would you present this?” are really “do you understand the deliverable is a decision?” — the candidate who leads with the “so what” and a recommendation beats the one with better statistics and a data dump
Related
- the course · databases interview bank (deeper SQL) · ai-ml interview bank (stats/DS)
- Interview Prep Index
Software engineering
Infrastructure & security
Computer science
Hardware & engineering