Data Analyst interview questions
54 questions asked in Data Analyst interviews, on SQL, dashboards, experiments and stakeholders. They are built from thousands of real job ads. Under each one: what the interviewer is listening for, and the follow-up they may ask. For the recruiter call, see the HR screen questions.
Behavioral questions
Questions about how you work in this role. More in behavioral questions for any tech role.
-
1
Tell me about a time your analysis changed a product or business decision.
All levels
Listening for A clear question, what they found and how it changed the decision.
Follow-up Ask how they knew the decision was right afterwards.
-
2
Tell me about a time you found an error in a report that people had been using for a while.
All levels
Listening for How they confirmed it, who they told and how they handled the fallout.
Follow-up Ask what decisions had been made on the wrong number.
-
3
Tell me about a time a stakeholder asked for one thing, but the real question turned out to be different.
Junior, Mid
Listening for Asking what decision the request supports before building.
Follow-up Ask what they ask stakeholders first now.
-
4
Tell me about a time you had to answer a question with data that wasn't good enough. How did you handle it?
Junior, Mid
Listening for Being clear about limits and still giving a useful answer.
Follow-up Ask how they explained the uncertainty.
-
5
Tell me about a time your recommendation was ignored. What did you do next?
Mid, Senior
Listening for Accepting the call or coming back with better evidence, without sulking.
Follow-up Ask what they would do differently in how they presented it.
-
6
Tell me about a time someone pushed you to present results so they looked better than they were.
Mid, Senior
Listening for Holding a line on honest numbers while staying helpful.
Follow-up Ask how the relationship went after that.
-
7
Tell me about a time you helped people outside the data team answer their own questions with data.
Mid, Senior
Listening for Tools, training or docs that really reduced requests.
Follow-up Ask how they measured that it worked.
Technical discussion questions
Talking through your experience and decisions, not a coding test. Grouped by topic.
SQL
-
1
Walk me through an analysis you did recently, from the question you were asked to what you found.
Junior
Listening for A clear question, the data used, steps taken and a result someone acted on.
Follow-up Ask what they would do differently now.
-
2
How do you check that a query's results are right before you share them?
Junior
Listening for Comparing with a known number, row counts, checking joins and filters, a few rows by hand.
Follow-up Ask about a time a check caught a mistake.
-
3
How do you raise the quality of analysis across a team of analysts?
Senior
Listening for Peer review, shared code and templates, standards that stuck.
Follow-up Ask about one change that made the biggest difference.
Dashboards
-
4
Yesterday's revenue dashboard shows half the usual number. How would you find out what happened?
Junior, Mid
Listening for Checking data freshness and pipelines before assuming a real drop, then telling users.
Follow-up Ask who they would tell and when.
-
5
Tell me about a dashboard you owned. How did you make sure people actually used it?
Mid
Listening for Working with users, fewer better metrics, checking usage, removing what nobody reads.
Follow-up Ask what they cut.
-
6
Analysts spend most of their time answering the same simple questions. How would you change that?
Senior
Listening for Self-serve tools on a clean shared model, training and docs.
Follow-up Ask how they would measure if it worked.
-
7
We have hundreds of dashboards, and nobody knows which ones are right. What would you do?
Senior
Listening for Usage data, certified dashboards with owners, removing the rest.
Follow-up Ask how they would handle someone who wants to keep an old one.
Metrics
-
8
How would you define an 'active user' for our mobile app?
Junior, Mid
Listening for A definition tied to real value, not just opening the app, and its trade-offs.
Follow-up Ask how the number changes with their definition.
-
9
Weekly active users dropped eight percent last week. How would you investigate?
Mid
Listening for Data checks first, then splitting by segment, platform and release, then outside causes.
Follow-up Ask when they would raise it with the team.
-
10
A team hit its target metric, but the product didn't really get better. How would you change the way targets are set?
Senior
Listening for Spotting proxy metrics that can be gamed, pairing goals with guardrails and checking with qualitative signals.
Follow-up Ask for a metric they have seen go wrong.
-
11
A new product line has no metrics yet. How would you decide what the team should track?
Senior
Listening for A main goal metric, input metrics and guardrails tied to the strategy.
Follow-up Ask how they would know a metric is wrong.
Python
-
12
When do you reach for Python instead of doing everything in SQL?
Junior
Listening for Clear reasons like stats, charts, messy files or automation.
Follow-up Ask about the last time they did that.
-
13
Tell me about an analysis that needed Python rather than a dashboard. What did you do in it?
Mid
Listening for A real case like modelling, forecasting or messy data, explained simply.
Follow-up Ask how they shared the result.
Stakeholder management
-
14
A manager asks you 'how are we doing?'. How would you turn that into something you can analyse?
Junior
Listening for Asking about goals and decisions, then choosing a few metrics.
Follow-up Ask which questions they would ask the manager.
-
15
How would you present a result to a manager who only has two minutes?
Junior
Listening for Answer first, one clear chart, what to do next.
Follow-up Ask what they would leave out.
-
16
A stakeholder wants a number by tomorrow, but the data needs a week of cleaning. What would you do?
Mid
Listening for A quick rough answer with clear caveats plus a plan for the proper one.
Follow-up Ask how they would word the caveat.
-
17
Five teams want analysis from you this quarter. How do you decide what to work on?
Senior
Listening for Ranking by impact and decisions at stake, being open about trade-offs.
Follow-up Ask how they say no to a senior leader.
Data visualization
-
18
Tell me about a chart you made that people found easy to understand. Why did you pick that type of chart?
Junior
Listening for Choosing the chart for the message and the audience.
Follow-up Ask what they removed to make it clearer.
-
19
A chart in a leadership deck makes a small change look huge. How would you redesign it?
Mid
Listening for Honest axes, context like past variation, clear labels.
Follow-up Ask how they would raise it with the person who made it.
-
20
How would you design a weekly dashboard for the leadership team?
Senior
Listening for A few key metrics, targets and trends, comments on changes, quick to read.
Follow-up Ask what they would leave off.
Data modeling
-
21
When you start working with a new set of tables, how do you figure out how they fit together?
Junior
Listening for Finding keys and grain, checking joins, reading docs or asking owners.
Follow-up Ask what one row means in the main table.
-
22
Two dashboards show different revenue numbers for the same month. How would you sort it out?
Junior, Mid
Listening for Tracing both back to their definitions and sources, agreeing one definition, fixing the model.
Follow-up Ask how they would stop it from happening again.
-
23
You join a startup as the first senior analyst, and the data sits in five tools with no warehouse. What would you set up first?
Senior
Listening for A simple warehouse, a few core models, the key metrics leaders need, ordered by value.
Follow-up Ask what they would avoid building early.
-
24
How would you design the core tables that most of the company's analysis is built on?
Senior
Listening for Clear grain, shared definitions, tests, docs and owners.
Follow-up Ask how they would change a core table without breaking reports.
Experimentation
-
25
An A/B test shows the new button got three percent more clicks. What would you check before saying it worked?
Junior
Listening for Sample size, significance, test length and whether clicks is the right metric.
Follow-up Ask what they would do if the result is not significant.
-
26
A product manager wants to stop a test after two days because it looks good. What would you tell them?
Junior
Listening for Why early peeks mislead, weekly patterns, the planned sample size, said kindly.
Follow-up Ask what they would do if the manager insists.
-
27
Product wants to test a new pricing page. How would you design the experiment?
Mid
Listening for Main metric, guardrails, sample size, test length and who sees which version.
Follow-up Ask how they would handle a test that needs two months.
-
28
A test shows more sign-ups but lower revenue per user. What would you recommend?
Mid
Listening for Looking at total value over time and segments, then a clear recommendation.
Follow-up Ask what more data they would want.
-
29
How would you set up an experimentation practice for a company that has only run a few tests?
Senior
Listening for Tooling, standards, review of designs, education and a record of results.
Follow-up Ask what they would do first.
Statistics
-
30
Tell me about a time you reported the median instead of the average. Why was it the better choice?
Junior
Listening for Skewed data and outliers explained with a real case.
Follow-up Ask how they would explain the choice to a manager.
-
31
A test result isn't statistically significant, but the product lead wants to ship anyway. How would you explain it?
Mid, Senior
Listening for Plain explanation of uncertainty and the risk either way, without blocking.
Follow-up Ask when shipping anyway is fine.
-
32
Conversion went up in every country but down overall. How would you explain that to the leadership team?
Senior
Listening for The mix of traffic shifting between countries, explained plainly with a simple chart.
Follow-up Ask which number they would tell leaders to watch.
-
33
Leadership wants a sales forecast for next year to plan hiring. How would you build it?
Senior
Listening for A simple model, key drivers, ranges instead of one number, regular updates.
Follow-up Ask how they would explain the uncertainty.
Data quality
-
34
Part of your data is missing for one week. How would you handle it in your analysis?
Junior
Listening for Finding out why, then excluding, filling or flagging it openly.
Follow-up Ask how they would tell readers about it.
-
35
How would you catch broken data before a stakeholder does?
Junior, Mid
Listening for Tests, freshness and volume alerts, checks on key metrics.
Follow-up Ask which alert they would set up first.
-
36
Leaders have stopped trusting the numbers from the data team. How would you win that trust back?
Senior
Listening for Finding why, fixing key metrics first, ownership, open communication about issues.
Follow-up Ask how they would know trust is back.
Tableau
-
37
Tell me about a dashboard you built in Tableau or a similar tool. Who used it, and for what?
Junior
Listening for Knowing the users and the decision it supports.
Follow-up Ask what they changed after people started using it.
-
38
A Tableau dashboard takes a minute to load, and people stopped using it. What would you do?
Junior, Mid
Listening for Fewer views and filters, pre-aggregated extracts, simpler layout.
Follow-up Ask how they would decide what to remove.
Product analytics
-
39
Fewer people are finishing sign-up this month. How would you find where they drop off?
Junior
Listening for Building the funnel step by step, splitting by device or source, checking tracking.
Follow-up Ask how they would tell a real drop from a tracking bug.
-
40
How would you show whether users who joined this year stay longer than those who joined last year?
Mid
Listening for Cohort retention curves with fair comparison windows.
Follow-up Ask what could make the comparison unfair.
Dbt
-
41
How has dbt, or a similar tool, changed the way you write SQL?
Junior, Mid
Listening for Reusable models, tests, docs and version control.
Follow-up Ask which test they find most useful.
Looker
-
42
In Looker, metrics are defined once in the model instead of in each chart. Why does that matter?
Junior, Mid
Listening for Consistency, one place to fix things and trust across dashboards.
Follow-up Ask about a time two dashboards disagreed.
Causal inference
-
43
Users who use our chat feature spend more money. Does that mean the chat feature makes people spend more?
Junior
Listening for Correlation versus causation and selection, like big spenders using chat more.
Follow-up Ask how they would test it properly.
-
44
Sales went up the week after we redesigned the checkout. What else could explain the increase?
Junior
Listening for Seasonality, campaigns, pricing, tracking changes, and how to rule them out.
Follow-up Ask how they would check the redesign's real effect.
-
45
Marketing wants to know if the welcome email makes people buy more. How would you find out?
Mid
Listening for A holdout group that gets no email, the right metric and time window.
Follow-up Ask how they would answer if a holdout is not allowed.
-
46
We launched a feature to everyone without a test. How would you estimate its impact?
Mid, Senior
Listening for Before-and-after with care, a comparison group, methods like difference-in-differences, limits stated.
Follow-up Ask what would make them distrust their own estimate.
-
47
Marketing says a TV campaign drove last month's growth. How would you check that claim?
Senior
Listening for Comparing regions or times, other causes, seasonality, being clear about limits.
Follow-up Ask how they would present a result marketing won't like.