25 Data Analyst Interview Questions With Sample Answers
25 data analyst interview questions career changers should prepare for, with clear answers on SQL, statistics, case studies, behavioral rounds, and a checklist.
Data analyst interviews test four things: SQL, statistics, business reasoning, and how clearly you explain a finding. SQL alone appears in 52.9% of data analyst job postings, so expect at least one live query question. The role is one of the fastest-growing in the U.S., with the Bureau of Labor Statistics projecting 34% growth for data scientists from 2024 to 2034.
You do not need to sound like a computer science graduate. You need to show that you can pull a correct answer out of messy data and explain it to someone who does not write code. This guide lists the 25 questions that come up most, grouped by type, with how to answer each one. For the broader career-change interview playbook, see interview prep for career changers.
What the data analyst interview looks likePermalink to “What the data analyst interview looks like”
Most data analyst interviews run three to five rounds. The structure is fairly consistent across companies.
| Round | What happens | What they test |
|---|---|---|
| Recruiter screen (20 to 30 min) | Background, salary, motivation | Communication, fit |
| SQL or take-home test | Write queries on a live dataset or solve a short assignment | Technical accuracy |
| Technical interview (45 to 60 min) | Live SQL, statistics, a small case | Depth of skill |
| Case or manager round | Business problem, metrics, stakeholder scenario | Business sense |
| Behavioral or team fit | STAR stories, collaboration | Communication, self-awareness |
Career changers usually struggle most in the technical and case rounds, where old job titles do not help. That is where to spend your prep time.
SQL questionsPermalink to “SQL questions”
SQL is the one skill you cannot bluff. It appears in 52.9% of data analyst job postings, more than any other language. Practice these until you can write them without notes.
1. What is the difference between INNER JOIN and LEFT JOIN? INNER JOIN returns only rows that match in both tables. LEFT JOIN returns every row from the left table plus any matches from the right, filling unmatched columns with NULL. Use LEFT JOIN when you must keep rows that have no match.
2. What is the difference between WHERE and HAVING? WHERE filters rows before grouping. HAVING filters after GROUP BY runs its aggregates. When you need to filter on a sum or count, use HAVING.
3. Write a query to find the second-highest value in a column.
Use a window function: SELECT value FROM (SELECT value, DENSE_RANK() OVER (ORDER BY value DESC) AS rnk FROM t) x WHERE rnk = 2;. DENSE_RANK handles ties without skipping ranks.
4. How do you find duplicate records?
Group by the identifying columns and filter: SELECT id, COUNT(*) FROM t GROUP BY id HAVING COUNT(*) > 1;.
5. What can a window function do that GROUP BY cannot? Window functions compute a result across rows without collapsing them. You can rank rows, build running totals, or compare each row to the previous one while keeping every row in the output.
6. What is a CTE, and when do you use it? A CTE is a named temporary result written in a WITH clause. Use it to break a complex query into readable steps and to avoid repeating the same subquery.
7. How do NULLs behave in SQL?
NULL means unknown, not zero. Comparisons with NULL return unknown, and most aggregates ignore NULL. Test with IS NULL, never with = NULL.
8. How would you make a slow query faster?
Read the execution plan, confirm the right indexes exist, filter early with WHERE, drop SELECT *, and reduce joins on unindexed columns.
Statistics and probability questionsPermalink to “Statistics and probability questions”
You do not need advanced math. You need to reason correctly about averages, uncertainty, and bias.
9. When would you use the median instead of the mean? When the data is skewed or has outliers. A few large values pull the mean up, while the median stays representative. Income and housing prices are the classic examples.
10. What is a p-value, and what is it not? A p-value is the probability of seeing results at least as extreme as yours if the null hypothesis were true. It is not the probability that your hypothesis is true, and a small p-value does not prove your effect is large.
11. How would you run an A/B test? Pick one metric and a success threshold before you start, split users randomly into control and variant groups, run it long enough to reach statistical power, then compare the groups and check the split for bias before you conclude anything.
12. What is selection bias, and how do you spot it? Selection bias happens when the sample is not representative of the population. Spot it by asking who is missing from the data and whether the missing group would change the conclusion.
Case study and business questionsPermalink to “Case study and business questions”
This round tests whether you can turn a vague business question into a data question. Talk through your reasoning out loud.
13. "A core metric dropped 20% overnight. How do you investigate?" First confirm the drop is real by checking the data pipeline, logging, and time zone. Then slice the metric by segment: geography, device, platform, and user type. Find the segment where the drop concentrates, then form a hypothesis.
14. "How would you measure the success of a new feature?" Name the primary metric the feature is meant to move, add two or three guardrail metrics that should not get worse, and explain how you would attribute any change to the feature rather than to something else.
15. "Estimate how many users do X per day." Break the estimate into pieces you can reason about, such as total users, the share who do X, and how often. State every assumption out loud and sanity-check the order of magnitude.
16. "Two metrics moved in opposite directions. What do you do?" Check whether they measure the same or different populations, watch for a composition effect (Simpson's paradox), and re-aggregate at a level where the relationship is clear.
Python, pandas, and data tool questionsPermalink to “Python, pandas, and data tool questions”
17. In pandas, what is the difference between merge and concat?
merge joins two DataFrames on shared columns or indexes, like a SQL join. concat stacks DataFrames along an axis without matching keys.
18. How do you handle missing data? First find out why values are missing. Then decide: drop them if they are few and random, fill them if you have a defensible value, or flag them as a separate category when the missingness itself carries information.
19. How do you choose which chart to use? Match the chart to the question. Line charts for change over time, bar charts for comparing categories, scatter plots for relationships, and histograms for distributions. Avoid pie charts once you have more than two or three slices.
Behavioral and career-change questionsPermalink to “Behavioral and career-change questions”
This is where career changers can win. Bring two or three STAR stories (Situation, Task, Action, Result), each one showing a different skill.
20. "Why are you moving into data analysis?" Frame the switch as moving toward data, not away from your old field. Name the part of your previous work that was already analytical, what you did to build on it, and where it leads. A concrete answer beats a passionate one.
When I switched from sysadmin work into analytics, the answer that moved me forward was not a clever SQL trick. It was walking the panel through a HealthTech dashboard where I had traced a spike in patient readmissions back to a timing bug in the data pipeline. The story showed I could find a real problem inside real data.
21. "Tell me about a data project you are proud of." Pick one portfolio project. State the business question, the data you used, the choice you made, and the result in numbers. For help building projects that hold up under this question, see best first projects for career changers into analytics and the complete portfolio guide for career changers.
22. "Describe a time you explained a technical finding to a non-technical stakeholder." Show that you turned a number into a decision. Name the audience, the simplification you made, and what the stakeholder did differently afterward.
23. "Tell me about a mistake you made with data." Own a specific mistake, explain what you learned, and describe the check you added so it would not happen again. Interviewers want self-awareness, not perfection.
Process and design questionsPermalink to “Process and design questions”
24. "How do you ensure data quality in your analysis?" Check for duplicates, missing values, and values outside plausible ranges. Cross-check totals against a trusted source, and document every transformation so someone else can reproduce your result.
25. "Walk me through a dashboard you would build for the sales team." Start from the decision the team needs to make. Lead with one headline metric, add two or three supporting breakdowns, and keep the layout to what fits one screen. Explain which filters matter and where you would set alert thresholds.
Three mistakes career changers make in data analyst interviewsPermalink to “Three mistakes career changers make in data analyst interviews”
- Apologizing for the career change. "I know I do not have a CS degree, but..." tells the interviewer to doubt you before you give evidence. State what you can do, then prove it with a project.
- Freezing on a live SQL question. If you do not know the exact syntax, say how you would approach it: which tables, which join, which filter. Clear thinking beats perfect syntax.
- Skipping the portfolio walkthrough. Saying "I learned SQL" is a claim. Walking through a query that found a real insight is proof. Not ready yet? Build projects first, starting with the SQL learning path for data analytics.
Data analyst interview prep checklistPermalink to “Data analyst interview prep checklist”
- Five SQL patterns you can write from memory: joins, GROUP BY and HAVING, window functions, CTEs, NULL handling
- One A/B test story you can explain end to end, including the metric and the sample split
- Two case frameworks: one for a metric drop, one for measuring a feature
- Three STAR stories: a project, a stakeholder explanation, a mistake you learned from
- One portfolio project you can walk through in under two minutes with a number in the result
- The job posting read closely, with three requirements mapped to your evidence
- Salary research done so you can answer the recruiter-screen number with confidence, using current data analyst salary data
Prepare with the work you already havePermalink to “Prepare with the work you already have”
You do not need to memorize 200 questions. You need five SQL patterns, three stories, and one project you can defend. That is enough to walk into a data analyst interview confident about what you bring and what it is worth: a national average near $82,640 a year, with most roles landing between $72,156 and $121,965 (Glassdoor).
How Traecta helps: it takes the data projects you have already shipped and maps each one to the exact skills the job posting lists, so you walk in with a talking point for every requirement instead of guessing what to highlight. If you want a structured path from where you are now to interview-ready, Traecta — Your Personalized Career Roadmap builds that plan from your current skills, your target role, and your timeline.
