Data Analyst Interview Questions

Harder than most candidates prepare for, because it is timed and against an unfamiliar schema. The recurring asks are cohort retention, funnel conversion by step, top-N per group using a window function, and running totals or period-over-period comparisons. Interviewers watch whether you clarify the schema before writing and whether you sanity-check your own output. Practising these five query shapes until they are automatic covers the large majority of what is asked.

Analysts describe and diagnose what happened; scientists predict and establish causality. Analyst work centres on SQL, BI tooling, metric definition, and stakeholder communication, with statistics used mostly for significance and sizing. Scientist work adds experimental design, causal inference, and modelling. In practice the boundary varies by company — at smaller organisations one analyst does both — but the interviews differ sharply, so target the loop you can pass.

The decision each analysis produced. "Identified that 31% of signup drop-off came from one verification step; removing it lifted completed signups 22%" is the shape that works, because it names the finding, the action, and the result. Also state the scale you worked at — rows, users, revenue covered — and the tools, but keep tooling brief. Listing dashboards built without saying what changed is the most common weakness in this discipline's CVs.

It strengthens your position but SQL is the non-negotiable one. Many analyst roles run entirely on SQL plus a BI tool, and interviews reflect that. Python becomes important when you want to move toward analytics engineering or data science, and it is often what distinguishes candidates for senior analyst roles where automation and reproducibility matter. If you have limited preparation time, get SQL to a high standard first.