Practice data analyst interviews out loud: metric investigations, SQL reasoning, stakeholder cases, and adaptive follow-ups, scored against your target level.
Opening question
“Daily active users dropped 12% this week. Walk me through your investigation.”
Analyst interviews test whether your reasoning is as structured as your queries. The classic prompt, "metric X dropped, investigate", has no right answer; it has a right SHAPE: hypotheses before queries, segments before conclusions, and a clear stopping rule. Interviewers push exactly where the shape gets fuzzy.
Illustrative examples. Practice sessions draw from calibrated question pools for your level.
Open with the hypothesis tree, out loud: instrumentation, seasonality/mix, then real change. It buys you structure points before you've analyzed anything.
For any query you'd write, rehearse the sentence "this would mislead me if…". That sentence is what interviewers are listening for.
End investigations with "so I'd do X, and I'd validate it by Y". Analysis that stops at the number reads as junior at every level above entry.
Pick Data Analyst and your level, then work investigations out loud with follow-ups that probe your hypothesis order, your query logic, and the recommendation you land on, scored per dimension against your target level's bar.
Feedback in this shape (sample)
The answer
“I'd query the funnel and see where the drop is.”
What it was missing
The hypothesis tree. Which failure class are you testing first, and what result would rule it out? Structure is the scored dimension here, not SQL.
Before the Data Analyst interview comes the screen that gets you into it. Build an ATS-ready resume free, no account needed to start.
Reading about interviews isn't practicing them. Take one and see where your answers stand. Start free, no credit card.
Photo: AI-generated editorial illustration
