The most commonly asked data analyst interview questions in India, with detailed model answers. Covers technical, behavioural, and situational questions asked by Indian recruiters.
Model answer
SELECT MAX(salary) FROM employees WHERE salary < (SELECT MAX(salary) FROM employees). Alternative with DENSE_RANK: SELECT salary FROM (SELECT salary, DENSE_RANK() OVER (ORDER BY salary DESC) AS rnk FROM employees) t WHERE rnk = 2. The window function approach handles ties correctly — if two employees share the highest salary, DENSE_RANK still returns a meaningful second.
Model answer
INNER JOIN returns rows where the join condition matches in BOTH tables. LEFT JOIN returns ALL rows from the left table and matching rows from the right (NULLs where no match). FULL OUTER JOIN returns all rows from both tables, with NULLs where no match on either side. Most analytical queries use LEFT JOINs to preserve all records from the primary dataset while enriching with dimension data.
Model answer
A dimension is a qualitative attribute used to slice/filter data — e.g., product category, geography, customer segment. A measure is a quantitative value that can be aggregated — e.g., revenue, units sold, session count. In a star schema, dimension tables surround the central fact table which contains measures and foreign keys to dimensions. This distinction drives dashboard design in Tableau and Power BI.
Model answer
Use STAR with these elements: What data you had access to, what question the business had, what analytical technique you used (cohort analysis, funnel analysis, regression, A/B test), what you found, how you communicated it (chart, dashboard, executive summary), and what decision was made. Quantify the outcome: "Led to a ₹40L cost reduction" or "Increased conversion by 12%".
Model answer
First understand WHY data is missing: Missing Completely at Random (MCAR), Missing at Random (MAR), or Missing Not at Random (MNAR). Options: (1) Drop rows if <5% missing and MCAR; (2) Mean/median imputation for numerical, mode for categorical; (3) Forward/backward fill for time-series; (4) Model-based imputation (KNN, MICE) for complex patterns; (5) Flag as a separate category for categorical. Never impute blindly — understand business context first.
Model answer
Never dismiss pushback — a stakeholder disputing data often knows something about the business context you don't. First, listen to their specific objection. Then: (1) Show your methodology transparently — data source, cleaning steps, aggregation logic; (2) Check if their counter-example reveals a data quality issue; (3) Run the analysis from their perspective and compare; (4) If the dispute is about interpretation not data, involve a neutral third party or senior stakeholder.
Use CV Prime to build an ATS-optimised Data Analyst CV tailored to the exact job description — so you pass the automated screen before the interview even happens.
CV Prime is a free CV maker and free AI CV builder for India. No credit card required.
Help us improve CV Prime
We use privacy-conscious product analytics only after consent. No CV text or API keys are tracked.