Role preparation
Data Analyst interview questions and answers.
SQL, statistical reasoning, and useful business evidence.
Questions
Technical questions to practise out loud
- 01
What is the difference between LEFT JOIN and INNER JOIN?
INNER JOIN returns only rows with a match on both sides; LEFT JOIN keeps every left-side row and fills unmatched right-side columns with NULL. Choose based on whether missing relationships must remain visible.
Read the full breakdown - 02
What does a p-value tell you?
A p-value is the probability, assuming the null hypothesis and model assumptions are true, of observing a result at least as extreme as the data. It is not the probability that the hypothesis is true.
Read the full breakdown - 03
When do you use SQL window functions?
Window functions calculate ranks, running totals, lags, or aggregates across related rows without collapsing them into one row. PARTITION BY defines the group and ORDER BY defines sequence within the window.
Read the full breakdown - 04
When do you use a SQL CTE?
Use a CTE to name a logical query step, express recursion, or make transformations readable. It is primarily a query-structure tool; whether it materializes and improves performance depends on the database and plan.
Read the full breakdown - 05
How do you define a reliable business metric?
A reliable metric has a documented business meaning, entity and event grain, numerator, denominator, filters, time zone, attribution window, exclusions, owner, and validation against a trusted source.
Read the full breakdown - 06
How do you validate data quality?
Validate data quality with automated checks for schema, completeness, uniqueness, validity, referential integrity, freshness, volume, and distribution drift, each tied to an owner and a response threshold.
Read the full breakdown - 07
How do you run cohort analysis?
Define cohorts by a stable entry event and period, choose an outcome such as retention, align elapsed time since entry, calculate comparable rates, and segment only when sample sizes remain meaningful.
Read the full breakdown - 08
How do you analyse a conversion funnel?
Define ordered funnel events, identity and conversion window first; compute users eligible and converted at each step, inspect biggest absolute and relative drop-offs, then segment and validate tracking before inferring causes.
Read the full breakdown - 09
How do you design a useful dashboard?
A useful dashboard serves one decision and audience, leads with a small set of defined metrics, shows target and trend context, supports intentional drill-down, states freshness, and avoids decorative charts without action.
Read the full breakdown - 10
How do you distinguish correlation from causation?
Correlation measures association; causation means changing one variable changes the outcome. Establish causality with randomization where possible or a defensible quasi-experimental design, temporal order, and control of confounders.
Read the full breakdown