Data analyst interviews usually test three things: can you get the data (SQL), can you reason about it (statistics), and can you turn it into a decision (business sense). Here are common questions in each, with worked answers.
SQL questions
Find the second highest salary in an employees table. One clean approach: SELECT MAX(salary) FROM employees WHERE salary < (SELECT MAX(salary) FROM employees); Mention that a window function like DENSE_RANK() handles ties and generalizes to the Nth highest.
A join doubled your revenue total. What happened? The joined table probably has duplicate keys, so each order matched more than one row. Check the join key for uniqueness and deduplicate before joining.
What does AVG do with NULLs? It skips them. Ratings of 5, 4, NULL and 3 average to 4, not 3. Say whether NULL should mean "missing" or "zero" for the business question.
Statistics questions
Mean or median for salaries? Median, when a few very high values pull the mean up. Five salaries of 70k, 72k, 75k, 78k and 300k have a mean of 119k but a median of 75k.
A test lifted conversion from 4.0% to 4.6%. Is it real? Ask about sample size first. With small samples that lift can be noise. Explain that you'd run a two proportion test or check the confidence interval, and look for a balanced split between groups.
What's the difference between percentage points and percent? Going from 4% to 5% is 1 point, or a 25% relative increase. Interviewers love this one.
Business questions
Signups dropped 10% this week. What do you check? Structure it: is the data right (tracking change, bot filter)? Is it real across all segments or one (device, country, channel)? What changed (release, campaign, seasonality)? Then size the impact.
Conversion fell but nothing broke. Check the mix. If mobile, which converts lower, grew from 60% to 75% of traffic, overall conversion drops even when each device is steady.
How to practice
Read answers once, then solve fresh problems under time pressure. See how to practice SQL for interviews for a weekly routine.
Questions
What SQL topics come up in data analyst interviews?
Joins, aggregation with GROUP BY, window functions, handling NULLs, and spotting duplicate rows.
Do data analysts need statistics?
Yes, the basics: mean vs median, distributions, A/B test significance and confidence intervals.
How should I answer a metric drop question?
Check data quality first, then segment the drop, then look for what changed, then size the impact.