Intervooh · Interview questions by job
Data Analyst interview questions (2026)
Researched, current questions asked in real data analyst interviews (Data & AI), with what a strong answer actually does. Questions marked 2026 are the newer, AI-era questions employers now ask.
Build my free day-by-day prep plan →
Tell it the company, role and date; it does the rest. Free, no card.
What they assess
- SQL & data manipulation
- Storytelling with data
- Business framing
- Data quality & rigour
- Stakeholder management
The questions to expect
How would you find the top five customers by revenue in each region? Talk me through the SQL.
This is a window-function question: ROW_NUMBER or RANK partitioned by region. Narrate the query shape before the syntax.
When would a LEFT JOIN and an INNER JOIN give you different numbers — and when has that bitten you?
Answer with a real duplicate-rows or dropped-rows story. Interviewers are checking you've debugged joins, not just defined them.
Sales dipped 12% last month. Walk me through exactly how you'd investigate.
Clarify the metric first, rule out data issues, then segment — by product, region, channel, new vs returning. A stated order is the skill.
Tell me about an insight you found that changed a decision. How did you present it?
One chart, one sentence of 'so what', one decision changed. Describe the presentation choices, not just the finding.
Describe a time you spotted an error in data everyone else trusted.
Show the habit that caught it — reconciling totals, sense-checking against a known figure — and how you raised it without blame.
Tell me about a stakeholder who asked for a report that wouldn't actually answer their question.
The move is asking 'what decision will this inform?'. Show you dug for the real question and delivered something more useful.
Why data analysis? What is it about working with data that suits you?
Connect a genuine trait — curiosity, pattern-spotting, tidy-mindedness — to a moment you felt it. Avoid 'I've always loved numbers'.
AI assistants can write your SQL now. What do you check before trusting what they produce?2026
Name checks: run it on a slice you know the answer for, inspect joins and filters, watch for silently dropped NULLs. Fluency plus scepticism wins.
Self-serve dashboards and AI Q&A tools let anyone pull numbers. Where does the analyst still add value?2026
Framing the question, defining metrics honestly, and catching wrong-but-plausible answers. Sound energised by this shift, not threatened.
Which visualisation would you pick to show a trend versus a comparison versus a distribution — and why?
Line for trend, bar for comparison, histogram for distribution — then earn points by naming a chart you'd refuse to use and why.
Walk me through a piece of analysis or a model that changed a business decision.
Name the decision, not just the deliverable. The strongest answers end with what the business did differently and what that was worth.
Tell me about a time a stakeholder challenged your numbers. How did you respond?
Show you checked before defending. Being openly willing to find your own error is what builds trust in your numbers.
Preparation notes
[object Object]
Turn this into a plan
A list of questions is a start; a programme is what changes the outcome. Intervooh builds a day-by-day plan for your exact data analyst interview — company research, story building with an AI coach, spoken practice with delivery feedback, and scored mock interviews.