Practice for
Data analyst interview practice
Short answer
Data analyst interviews mix technical questions you must explain in plain words (joins, correlation, A/B tests) with questions about turning numbers into decisions. Below is every data analyst question in the AnswerDrill bank with what a good answer covers.
What interviewers look for
- A repeatable method: how you check, clean and question data before trusting it.
- Plain explanations of technical ideas for people who do not use SQL.
- Business sense: which metric matters and what decision it supports.
- Honesty about limitations in the data.
Every data analyst question in the bank, with how to prepare
These are the technical questions AnswerDrill asks for this role, with the levels each one is asked at. A technical session picks up to four of them for your level and adds general questions that name your role.
How do you clean a messy dataset before analysis? Walk me through your steps. All levels
Go step by step: understand the source, check types and ranges, handle duplicates and missing values, standardise formats, and document every change so the work can be repeated.
Explain the difference between correlation and causation with an example. All levels
Use a simple example where a third factor drives both variables, then explain how an experiment or careful controls would test causation.
A key metric dropped 20% overnight. How would you investigate? All levels
Check data quality first (tracking or pipeline change), then segment by time, platform, region and channel, and look for releases or outside events. Say how you would report early findings.
Explain an inner join and a left join to someone who has never used SQL. Entry
Use two short lists, such as customers and orders: an inner join keeps only matches, a left join keeps every customer and leaves gaps where there is no order.
How do you choose the right chart for a piece of data? All levels
Start from the question: trends over time suit lines, comparisons suit bars, parts of a whole need care. Mention labelling, a clear title and avoiding clutter.
How do you present findings to people who are not technical? All levels
Lead with the answer and the decision it supports, show one clear chart, and keep methods in an appendix. Give an example of a presentation that changed a decision.
How do you handle missing or unreliable data? All levels
Find out why it is missing, measure how much, then choose: exclude, fill in with a stated method, or flag it. Always state the limitation in your findings.
How would you design an A/B test for a new checkout button? Mid and senior
Define the hypothesis and main metric, split users randomly, estimate the sample size needed, run for full weeks, and check guardrail metrics before deciding.
Tell me how you would build a dashboard for a sales team from scratch. Mid and senior
Start by interviewing users about the decisions they make, pick a few metrics that drive those decisions, agree definitions, then build, test with users and iterate.
How would you set up data quality checks for a reporting pipeline? Senior
Mention automated checks for freshness, volume, nulls, duplicates and value ranges, alerts that reach an owner, and how you would handle a failed check before reports go out.
Behavioural questions for this role
Prepare a story where your analysis changed a decision, one where the data was wrong and you caught it, and one where you explained a result to a sceptical audience. See the STAR method and common behavioural questions.
How to practise this role in AnswerDrill
- Open the practice app, choose the role, your level and "Technical".
- Answer each question out loud, up to two minutes by default, or type it.
- Read the scores, your filler words, your pace and the stronger answer, then answer your weakest question again.
Answers are spoken, so practise describing a query or a chart in words; AnswerDrill does not run SQL or read your charts.
Practise it out loud. Answer five questions for your role and get scores, tips and your filler words. 2 free AI sessions a week, no sign-up; your history stays on your device.
Questions
Do I need to write SQL in the practice?
No. The questions ask you to explain your approach out loud, the way many first-round interviews do.
Which level should I choose?
Entry level focuses on fundamentals such as joins; mid and senior add experiment design and dashboards; senior adds data quality checks.