Data analytics
Data Analyst Interview Prep When Your Job Title Has Nothing to Do with Data
Turn your existing work into credible data analyst interview stories, build a small portfolio and explain your career switch without overselling it.
How to talk about analysis when your old job title wasn’t “data analyst”
Walking into an interview with only a course certificate and saying “I’m a quick learner” gives the interviewer very little to work with. They still can’t tell how you handle messy records, check a number or explain what your result means.
Your previous job probably gives you better material than you think. If you’ve chased missing figures, compared targets with actual results, cleaned a register or explained a confusing report to a colleague, you already have examples worth discussing. The job is to find those examples, back them up with one small project and practise explaining what you actually did. No sudden promotion to “data strategy lead” required.
Start with questions, not your old designation
Job titles hide a lot of useful work. “Operations executive” could cover purchase approvals, delayed-order checks, vendor records and a weekly review. Only some of that belongs in your analyst story, so look for the times you used information to make or support a decision.
Go through the last few months of your work and write down four things:
- A recurring question someone asked you, such as “Why are these cases still pending?”
- The records you used to answer it
- What you checked, calculated or compared
- What changed after your answer, even if the change was small
That fourth point is where many answers fall flat. “Made reports” is a task. “Prepared the Monday pending-cases view so the team could decide which cases to follow up first” tells the interviewer why the report existed.
Say you worked in customer support and noticed lots of password-reset tickets. Don’t turn that into “built a company-wide retention strategy”. A version you can defend is narrower: you grouped ticket notes by issue, spotted repeated reset problems and raised the pattern with the relevant team. That gives the interviewer useful follow-ups. Which notes did you group? What counted as a reset issue? Who did you tell? You can answer without inventing a second career on the spot.
And if your work genuinely involved no analysis, say so. Use a portfolio project to show what you can do now. Your old job can still supply context around deadlines, messy inputs and explaining things to colleagues who don’t want a lecture about SQL.
Turn each example into a story you can defend
A useful interview story has a clear chain: question, records, method, decision. It doesn’t need a dramatic rescue or a suspiciously perfect result. Sometimes the turning point is that the export had duplicate rows and everyone was fed up with sorting it by hand.
Here’s an ordinary admin example shaped into an answer:
“Our team reviewed pending requests every week, but the spreadsheet mixed new and old entries, so the review often started with manual sorting. I checked how dates and statuses were recorded, removed confirmed duplicates with the owner of the sheet and created a view grouped by age and status. In the next review, we could start with the oldest open requests. My part was cleaning and organising the view; I didn’t change the approval process.”
That answer names the problem, the actions and the boundary of the person’s responsibility. It also invites sensible follow-ups. How did you identify duplicates? Who confirmed them? What counted as an old request?
Those are much better questions than “Can you explain this impressive number on your CV that you can’t reconstruct?”
Use “we” for the team’s outcome and “I” for your own contribution. That helps the interviewer separate your work from what the whole department achieved. It also stops a small improvement from sounding like a one-person transformation programme.
Prepare a few stories instead of forcing one tale to answer every question. One can cover cleaning and checking. Another can show analysis. A third can cover a mistake you caught or a conclusion you revised.
Yes, talking about a mistake is allowed. “I initially counted cancelled orders, then changed the filter after checking the status definitions” sounds like someone who can be trusted with real data. “I always make sure the data is accurate” sounds nice and tells the interviewer almost nothing.
A common trap here is using a metric without knowing its rule. If your dashboard says 420 open requests, be ready to explain whether that includes reopened requests, requests with blank statuses or requests closed after the export. You don’t need to recite the whole spreadsheet. You do need to know what went into the number.
Build one project that backs up your words
Your portfolio doesn’t need a row of flashy dashboards. One project with a clear question and visible working is more useful because you can discuss it properly.
Choose a question you understand. A retail professional might compare completed and cancelled orders. Someone from education might examine attendance patterns. Use public data, synthetic data or records you’re explicitly allowed to share.
Please don’t upload your employer’s spreadsheet and discover confidentiality in the middle of an interview.
Keep the folder easy to inspect. Include the question, a short data dictionary, your cleaning steps, the analysis and a brief write-up of what the results do and don’t show. If you changed a category or removed rows, record the rule. A chart with no explanation is just a colourful rectangle asking for trouble.
For a PostgreSQL order-analysis project, a first query might look like this:
SELECT
DATE_TRUNC('month', order_date) AS month,
SUM(amount) AS gross_order_value
FROM orders
WHERE status = 'completed'
GROUP BY 1
ORDER BY 1;
The interesting bit is the filter, not the fact that the query has nice indentation. Why count only completed orders? Where did that status definition come from? Are refunds handled elsewhere?
This is also where copying from a tutorial can bite you. In PostgreSQL, writing WHERE status = "completed" can produce ERROR: column "completed" does not existbecause double quotes are used for identifiers. The text value belongs in single quotes, as in the query above. If you get that error, check the quotes before assuming the column is missing.
PostgreSQL’s SELECT documentation and date and time function documentation are useful when you want to check the syntax instead of running a copied query and hoping it behaves.
Then add one honest limitation. Perhaps the sample covers too short a period. Perhaps some categories are ambiguous. Or perhaps order value doesn’t equal money received, which matters if someone starts calling your total “revenue”.
Beginners often hide limitations because they think it makes the project look weak. It usually does the opposite. It shows you know where your answer stops. A monthly total can be calculated correctly and still be the wrong number for the question being asked.
Finally, rehearse a short walkthrough. Open the folder, state the question, show one cleaning decision, run or explain one analysis step and finish with the limitation. Don’t spend ten minutes scrolling through every chart while the interviewer quietly ages.
We’d also keep the raw data unchanged and do cleaning in a separate step or file. If the interviewer asks why a row disappeared, you can then point to the rule instead of scrolling around and hoping memory saves you.
Practise the awkward questions directly
“Why should we consider you without analyst experience?” Try this shape:
“My previous title wasn’t data analyst, so I won’t present it as one. In my operations work, I regularly checked records and prepared views used in weekly decisions. I’ve since built a small order-analysis project where I documented cleaning, wrote the queries and explained the limits. Here’s what I did myself and what I’m still practising.”
For “Tell me about a challenge,” avoid making the whole answer about how emotionally difficult the career switch feels. That may be completely true, but the interviewer needs evidence of your working method. Talk about inconsistent statuses, missing dates, vague requirements or a chart people misunderstood.
For instance, what did you do when one sheet used Closedanother used closedand a third had blanks? Did you standardise the values? Did you check whether blanks meant unknown or not yet updated? That’s a work problem with decisions in it, which makes it useful interview material.
For a technical question you can’t answer, don’t invent syntax live on stage. Say what you know, describe how you’d test a solution and admit the missing piece.
You could add, “I haven’t used window functions in a project yet, but I’d first check whether a grouped query answers the question.” Clear and bounded beats confident nonsense.
Do at least one practice session aloud with another person. Ask them to interrupt with “Why?” after any step. Why that filter? Why that chart? Why remove those rows? If your answer is always “the tutorial did it”, you’ve found exactly what to revise.
This feels awkward. Do it anyway. Reading your answer silently is much easier than explaining why gross_order_value excludes cancelled orders while someone watches you.
Use the interview to check the role too
Ask what records a new analyst would handle, who checks their work and how questions reach the team. The answers help you judge whether the role involves actual analysis, endless manual reporting or a strange mixture of everything with “dashboard” sprinkled on top.
If nobody reviews a new analyst’s work, ask how they’re expected to learn the data definitions. If every request arrives as an urgent message with no clear question, expect to spend time untangling requests before touching a query. Those details matter more than whether the job listing uses the exact title you want.
Before the interview, keep one page with your story prompts, project question and two limitations you’d mention. Read that instead of memorising full speeches. You’ll sound like yourself, which is handy because you’re the person attending.
If you want guided practice, Rise Institute’s data analytics training focuses on exactly these skills: cleaning messy data, writing queries you can explain and defending your results out loud in mock interviews. Either way, your next useful step is to draft one work story and ask a friend to challenge every claim in it.
3 questions, about a minute
Check your understanding
- Data analytics
- Interview preparation
- Career switch
- Portfolio
- SQL
