Domain Deep-Dives · 9 min read
Data & Analytics Interview Deep-Dive: SQL, Stats & Product Cases
The three rounds that decide data analyst and data scientist offers.
Data interviews reward fluency across three distinct muscles: writing correct SQL fast, reasoning about statistics and experiments, and turning a vague business question into a measurable analysis. Strong candidates practice all three, not just the one they're most comfortable with. Here's how each round works and what interviewers reward.
SQL patterns worth drilling
- •Window functions: RANK/DENSE_RANK, ROW_NUMBER, LAG/LEAD, and running totals — the highest-frequency advanced topic.
- •Joins and their edge cases, especially self-joins and anti-joins (finding what's missing).
- •Aggregation with GROUP BY and HAVING, and when a subquery or CTE is cleaner than a join.
- •Nth-highest, top-N-per-group, and time-based cohort queries — classic prompts.
Statistics and A/B testing
Product-analytics cases
'How would you measure the success of feature X?' or 'engagement dropped 8% — investigate.' Structure before analyzing: define the goal and metric, segment the data, form hypotheses, and say how you'd validate each. This mirrors the PM metric-diagnosis approach and is very learnable.
- •Define a north-star metric plus guardrails for any feature you're asked to measure.
- •When diagnosing a drop, check data quality first — broken logging looks like a real change.
- •Segment aggressively: platform, geography, cohort, new vs existing.
Behavioral
Data roles are judged on communication as much as technique. Expect 'tell me about an analysis that changed a decision.' Use STAR, and emphasize how you translated a technical insight into a business action a non-technical stakeholder acted on.
Practice a technical mock
Rehearse SQL reasoning and analytics cases out loud with instant feedback.
Practice a technical mock →