Ace 
By the Ace Coach team, built with the best speech coaches · Updated
What these interviews test
O*NET's profile for data scientists, the closest match for analyst work, lists cleaning and manipulating raw data, creating visualizations, identifying business problems that data can address and delivering presentations of results to management. Interviews check each of those: a technical screen, a case and a presentation of findings.
The part candidates underestimate is the last step. Interviewers want a recommendation, not a pile of charts. Practice stating the answer first, then the evidence, in plain words.
9 questions and how to answer them
1. “How would you handle a dataset with missing or inconsistent values?”
What they're checking: Practical data cleaning judgment.
How to answer: First say you'd check how much is missing and why; then describe the choice (drop, fill, flag) and say you'd document it.
2. “Walk me through an analysis that changed a decision.”
What they're checking: Impact, not just technique.
How to answer: State the decision first, then the question, the method in one sentence, the finding and what the team did.
3. “Signups dropped 15% last week. How do you investigate?”
What they're checking: Structured thinking about causes.
How to answer: Check for data errors first, then split by channel, device and region, and compare with the same week last year.
4. “Explain a p-value to a manager.”
What they're checking: Plain-language explanation of statistics.
How to answer: Use a coin or A/B example and say what it does not tell you, then state the practical meaning.
5. “How do you choose the right chart?”
What they're checking: Communication judgment.
How to answer: Start from the question: trend over time is a line, comparison is a bar. Mention removing clutter.
6. “A stakeholder disagrees with your numbers. What do you do?”
What they're checking: Handling pushback with evidence.
How to answer: Ask what they expected, check definitions together and show your calculation step by step. Full answer guide
7. “Tell me about a mistake in an analysis.”
What they're checking: Honesty and checking habits.
How to answer: Describe how you caught it, who you told and the check you added, such as comparing totals to a source. Full answer guide
8. “How do you prioritize several requests from different teams?”
What they're checking: Judgment under load.
How to answer: Rank by decision value and deadline and tell teams what will be late. Full answer guide
9. “Tell me about yourself.”
What they're checking: A clear path into analytics.
How to answer: Lead with a result such as a report that saved hours, and why you're drawn to this team's data. Full answer guide
How to prepare
- Rehearse a findings talk. Give a two-minute answer, recommendation first, from a project in your portfolio.
- Refresh core SQL out loud. Joins, grouping and window functions, explaining each step as you write.
- Prepare a messy-data story. What was wrong, how you found it and what you changed.
- Learn their metrics. Find the three numbers the company reports publicly and think through what moves them.
Then say your answers out loud. In an AI mock interview the interviewer asks your role's questions and follows up, and you see where you rambled, rushed or hedged.
AI mock interview practice
Presentation practice
Free filler word counter