Data Analyst Interview Questions & Answers

SQL Basics

1. How do you filter rows in SQL?
Use a WHERE clause. It limits rows before grouping or sorting, which keeps the result focused.
Model answer:
SELECT * FROM orders WHERE order_status = 'complete';

2. How do you remove duplicate rows from a result set?
Use DISTINCT when you want unique values in the output. It works best for simple de-duplication needs.
Model answer:
SELECT DISTINCT customer_id FROM orders;

3. How do you join two tables in SQL?
Use a JOIN with matching keys. Start by identifying the common column, then choose the join type you need.
Model answer framework:

• Identify the shared key
• Pick INNER JOIN, LEFT JOIN, or another join
• Select only the fields you need

4. What is the difference between INNER JOIN and LEFT JOIN?
An INNER JOIN returns only matching rows. A LEFT JOIN returns all rows from the left table and matches from the right table when they exist.
Model answer:
Use INNER JOIN for matched records only. Use LEFT JOIN when you want to keep every row from the first table.

SQL Aggregation

5. How do you calculate total sales by month?
Use GROUP BY with an aggregate like SUM. That lets you roll transaction data into monthly totals.
Model answer framework:

• Extract the month from the date column
• Group by month
• Sum the sales amount

6. How do you count records in a table?
Use COUNT(). It helps you measure volume, spot gaps, and compare groups.
Model answer:
SELECT COUNT(*) FROM users;

7. How do you find the average order value?
Use AVG() on the order amount column. It gives a simple summary of typical purchase size.
Model answer:
SELECT AVG(order_amount) FROM orders;

8. How do you filter grouped results?
Use HAVING. It filters groups after aggregation, unlike WHERE, which filters rows before grouping.
Model answer framework:

• Group the data
• Aggregate the values
• Use HAVING for group-level filters

Data Cleaning

9. How do you handle missing values in a dataset?
First, identify where the gaps appear. Then decide whether to remove, fill, or flag them based on the analysis goal.
Model answer framework:

• Check which columns have missing values
• Assess how much data is missing
• Choose a treatment method that fits the use case

10. How do you find duplicate records?
Compare rows across key columns that should be unique. SQL can help you count repeated combinations.
Model answer:
Look for repeated values in fields like email, user ID, or transaction ID.

11. How do you standardize text values?
Use consistent casing, trimming, and formatting rules. This helps prevent issues like USA, usa, and Usa being treated as different values.
Model answer framework:

• Convert text to one case
• Remove extra spaces
• Clean inconsistent labels

12. Why does data cleaning matter in analysis?
Dirty data leads to wrong metrics and weak conclusions. Clean inputs make your findings easier to trust and explain.
Model answer:
Cleaner data supports better analysis, especially when the dataset comes from multiple sources.

Metrics and Business Thinking

13. How would you explain a key metric to a non-technical stakeholder?
Use plain language and connect the metric to a business outcome. Avoid jargon unless you define it.
Model answer framework:

• Name the metric
• Explain what it measures
• Say why it matters

14. What is the difference between a metric and a KPI?
A metric measures something. A KPI is a metric tied to a business goal.
Model answer:
All KPIs are metrics, but not all metrics are KPIs.

15. How would you choose the right metric for a problem?
Start with the question you want to answer. Then pick a metric that reflects that outcome clearly and consistently.
Model answer framework:

• Define the business question
• Identify the behavior or outcome to measure
• Confirm the metric is easy to track

16. How do you spot a misleading metric?
Check whether the metric hides context, mixes different groups, or changes because of data quality issues.
Model answer:
A number can look good while the underlying trend is weak, so always check the full picture.

Analytical Reasoning

17. How would you approach a sudden drop in conversions?
Break the funnel into steps and isolate where the drop starts. Then compare time periods, segments, and traffic sources.
Model answer framework:

• Confirm the drop is real
• Find the funnel step with the biggest change
• Segment the data to narrow the cause

18. How do you answer an open-ended analysis question?
Start with the goal, then define the data you need, and finally explain your conclusion clearly.
Model answer:
Use a structured approach. State the question, inspect the data, and present the most likely answer with evidence.

19. What would you do if two charts seem to tell different stories?
Check the time range, filters, and definitions used in each chart. Different setups often explain the mismatch.
Model answer framework:

• Compare the data sources
• Check date ranges and filters
• Verify the business definition behind each chart

20. How do you prioritize questions during analysis?
Focus on the question that changes the decision first. That keeps the work practical and avoids analysis that does not lead anywhere.
Model answer:
Start with the most decision-relevant question, then move to supporting details.

Beginner and Intermediate Practice Prompts

21. Write a SQL query to find the top 5 highest-value orders.
Use ORDER BY and LIMIT. Sort the order values from highest to lowest, then return the top 5 rows.
Model answer framework:

• Select the order table
• Sort by order value descending
• Limit to 5 rows

22. How would you calculate retention for a cohort?
Group users by signup period, then track how many return in later periods. The exact setup depends on the time window you choose.
Model answer framework:

• Define the cohort start date
• Track repeat activity over time
• Compare returning users to the original cohort size

23. You notice a spike in one dashboard metric. What do you check first?
Check for data changes, filter changes, and timing issues before assuming the business changed.
Model answer:
Start with the data pipeline and dashboard settings, then validate the source tables.

24. Explain how you would summarize analysis findings to a manager.
Lead with the main answer, then share supporting evidence and the likely business impact. Keep it short and direct.
Model answer framework:

• State the conclusion first
• Share 2 or 3 supporting points
• End with the takeaway for the business

Related tools

• Resume Analyzer
• ATS Resume Checker
• Resume Builder
• Job-Description Matcher

Turn your resume keywords into interview questions

For each term on your resume, prepare a question that tests actual experience. For SQL: “What grain did the output have, and how did you prevent joins from duplicating totals?” For dashboards: “How was the metric defined, and what happened when the source changed?” For cohort analysis: “Which event started the cohort and which return behavior did you measure?”

Prepare an answer using your actual project, your contribution, a check you performed, and a limitation. A classroom analysis is valid evidence when labelled accurately. Do not turn it into a claim that a company adopted your recommendation.

The product analytics keyword guide shows how metric definitions affect wording. Microsoft's star-schema reference provides useful terminology for explaining reporting grain and relationships.