Will I get a SQL test?

Usually, either as a live exercise or a short take-home. Expect joins, aggregation and at least one window function.

Do I need Python or R?

Only if the advert asks for it. Many analyst roles are SQL plus a BI tool; check which tools are named in the job description.

How do I prepare a portfolio?

One project explained end to end — the question, the data, the analysis, the decision it changed — is worth more than several dashboards.

What is the most common reason candidates fail?

Answering the question asked rather than the decision behind it, and presenting results without a recommendation.

Data analyst interview questions and preparation

A data analyst interview usually has three parts: a SQL or technical test, a case question about measuring something, and stakeholder questions about how you turn an ambiguous request into an answer people act on.

Written by Ace Editorial · Published 23 September 2026 · Last updated 23 September 2026

What is tested in a data analyst interview?

SQL fluency, analytical reasoning, how you define metrics, and how you communicate findings to people who are not analysts.

SQL — joins, aggregation, window functions, deduplication
Metric definition — what to measure and why
Diagnosis — "this number moved, find out why"
Communication — explaining a result to a non-technical stakeholder

What SQL comes up most often?

Joins with awareness of row duplication, group-by aggregation, window functions for ranking and running totals, and date handling.

Say your assumptions out loud: whether the join could fan out, whether nulls matter, whether the date column is event time or load time. Interviewers score that reasoning.

How do you answer "how would you measure X?"

Name the decision the metric supports, propose a primary metric, add a guardrail metric, then say what would make you distrust the number.

Example

"For onboarding I'd use the share of sign-ups completing a first meaningful action within seven days, with a guardrail on support contacts. I'd distrust it if a product change moved the definition of that first action, so I'd keep the raw funnel alongside it."

How do you answer stakeholder questions?

Show that you clarify the underlying decision before writing any query.

Example

"When someone asks for 'all the data on churn', I ask what they're going to decide with it. Nine times out of ten that turns a three-day extract into a one-chart answer."

What behavioural questions come up?

An analysis that changed a decision, a time your numbers were challenged, and a time you got something wrong.

What questions should you ask?

Ask about the data stack, how trusted the data currently is, who consumes the analysis, and how often the team's work changes a decision.

Frequently asked questions

Will I get a SQL test?: Usually, either as a live exercise or a short take-home. Expect joins, aggregation and at least one window function.
Do I need Python or R?: Only if the advert asks for it. Many analyst roles are SQL plus a BI tool; check which tools are named in the job description.
How do I prepare a portfolio?: One project explained end to end — the question, the data, the analysis, the decision it changed — is worth more than several dashboards.
What is the most common reason candidates fail?: Answering the question asked rather than the decision behind it, and presenting results without a recommendation.

Preparing for a data interview?

Ace reads the job description and your CV and shows the technical, case and stakeholder questions you're most likely to face — with the projects from your own history to use.

Build my interview pack

Free analysis first. You only pay once you've seen what we've found.