Screening a data analyst comes down to whether they can get data with SQL, check it before trusting it, and explain what it means to someone who doesn't work with data. Fifteen to twenty minutes is enough. The best signal is a story about an analysis that changed what a team did.
What does a data analyst do?
A data analyst answers business questions with data: how many customers came back, which region is slipping, whether a change worked. They pull data with SQL, analyze it in a spreadsheet or in Python or R, build dashboards in tools such as Tableau or Power BI, and present what they found. The nearest Bureau of Labor Statistics occupation is data scientists, who "use analytical tools and techniques to extract meaningful insights from data," with duties that include collecting and analyzing data and using data visualization software to present findings.
Which skills should you check at each level?
| Level | What they should be able to talk about |
|---|---|
| Junior (0 to 2 years) | SQL basics (filtering, grouping, joins), Excel (pivot tables, lookups), one dashboard tool, cleaning messy data, explaining a chart |
| Mid-level (2 to 5 years) | Harder SQL (window functions, common table expressions), building and maintaining dashboards, defining metrics with stakeholders, A/B test basics, Python or R, checking data quality |
| Senior (5+ years) | Owning metric definitions across teams, designing experiments, forecasting, turning a vague question into an analysis plan, presenting to leadership, mentoring |
What screening questions should you ask a data analyst?
We wrote these for this page. Swap in the client's tools and business.
Tell me about an analysis that changed a decision. What was the question, and what happened?
We'd ask this one first. It shows whether their work has ever led to action.
Strong answer: a specific question (why did renewals drop in one region?), the data they used, what they found, what the team changed, and roughly what happened next.
Weak answer: "I made weekly dashboards for management," with no decision attached.
How would you find each customer's most recent order?
This is a SQL check you can judge without writing SQL yourself.
Strong answer: describes grouping orders by customer and taking the latest date, or using a window function (they may say ROW_NUMBER or "partition by customer"), and mentions joining back to get the order details or handling two orders on the same day.
Weak answer: "I'd export the data to Excel and sort it."
The dashboard shows revenue fell sharply yesterday. What do you check before telling anyone?
Good analysts doubt surprising numbers first.
Strong answer: whether yesterday's data finished loading, whether a filter or time zone changed, duplicates or missing records, a comparison with the source system, and only then a breakdown by segment to find the real drop.
Weak answer: sends the alert straight to leadership.
A manager asks you for "a dashboard with everything." What do you do?
This shows whether they can shape a vague request.
Strong answer: asks what decisions the dashboard should support and who will use it, agrees on a few metrics and how each is defined, builds a small first version and adjusts it after feedback.
Weak answer: builds thirty charts and hopes some are useful.
How do you explain a finding to someone who doesn't like numbers?
This tests communication, which the hiring manager will judge in every meeting.
Strong answer: leads with the answer and what to do about it, uses one clear chart, says how sure they are, and keeps the method for anyone who asks.
Weak answer: walks through every step of the analysis in order.
How would you tell whether a website change really increased sign-ups?
Ask this of mid-level and senior candidates; it tests whether they separate cause from coincidence.
Strong answer: an A/B test with a control group if possible, enough visitors and enough time to trust the result, and a check for seasonality or other changes that happened at the same time. They explain significance in plain words.
Weak answer: "Compare this week with last week."
What do you do when two teams define the same metric differently?
Mismatched definitions are a common reason two reports disagree.
Strong answer: writes down both definitions, shows exactly where they differ (for example, "active customer" meaning logged in versus bought something), gets the metric's owners to agree, and labels dashboards clearly.
Weak answer: quietly picks one and moves on.
What red flags should you note?
- SQL, Python, R, Tableau, Power BI and SAS on the resume, but every example is in Excel.
- They can't explain what a join does in plain words.
- Every story is about tools and none is about a business question.
- Numbers in their stories change when you ask about them a second time.
- Machine learning talk for a reporting role, with nothing on how their work was used.
A red flag is a reason to ask more in your own call. If you suspect someone else is answering, see how to spot a proxy interview.
How do you run this screen with Simplehire?
Create the role with the client's tools and needs as its required skills, for example SQL, Power BI, Excel, A/B testing and stakeholder communication. The candidate talks with our AI interviewer in the browser, and you get a screening report as soon as they finish.
For junior and high-volume analyst roles, Basic ($2 per screen, about 15 minutes) is usually enough to find the people worth a call. Before you submit someone, Adaptive ($3, about 20 minutes) is the better fit: follow-up questions test whether an analysis story holds up, and every answer gets an AI assessment. Senior or client-critical roles fit Advanced ($5, about 30 minutes), which goes deeper on strong answers and adds an optional coding round, scored separately, for analysts who write code.
Read Strengths next to Areas to probe. If the strengths are all tools and the open questions are about explaining results, spend your own call on how they work with stakeholders. The verdict is a recommendation, and who goes forward is your call. Proctoring runs on every screen, and 25 Basic screens a month cost nothing (pricing).
Related pages: data engineer questions, business analyst questions and Python developer questions.
Frequently asked questions
What skills should a data analyst have?
SQL first, then a spreadsheet tool such as Excel and a dashboard tool such as Tableau, Power BI or Looker. Mid-level and senior analysts often use Python or R as well. Just as important is the ability to turn a vague business question into a specific one and explain the answer in plain words.
How can a non-technical recruiter test SQL in a screen?
Ask the candidate to explain, in words, how they would answer a simple question with a query, such as finding each customer's most recent order. Listen for named steps and terms such as joins, grouping and window functions. A candidate who only says they'd export to Excel hasn't shown SQL.
Is a data analyst the same as a data scientist?
They overlap. Analysts mostly describe what happened and why, using SQL, spreadsheets and dashboards. Data scientists more often build statistical or machine learning models. Job titles vary between companies, so read the job description.
