Data Analyst Interview Questions

What should you focus on for a Data Analyst interview?

Data quality, SQL query writing and data pipelines appeared more often in these Data Analyst interviews than across all interviews. Questions leaned toward mid-level and junior work: 48% were mid-level and 31% were junior. The checks below cover explaining your choices as well as doing the work.

Can you show how you checked and fixed data?

Questions covered missing data, duplicates, validation and comparing records across sources. They also asked how candidates checked an unfamiliar dataset before trusting it. Check whether your answer covers both the change you made and how you verified it.

• Can you describe a data issue you found, what you changed and how you checked the fix?
• For an unfamiliar dataset, which checks would you use for missing values, duplicates, completeness and accuracy?
• If records differ across sources, how would you compare them and investigate the differences?

Can you write SQL, not just explain it?

SQL joins were the most asked question, appearing in 8 of 129 interviews. Code-writing tasks appeared in 23 of 129 interviews. Check both your explanations and your queries.

• Which rows do INNER, LEFT, RIGHT and FULL OUTER JOIN keep when there is no match?
• Can you explain the difference between WHERE and HAVING, and between RANK and DENSE_RANK?
• Can you write a query to find the highest-paid employee in each department?
• Can you identify duplicate records in a SQL table?
• When would you use a CTE rather than a subquery?

Can you explain how data reaches a report?

Pipeline questions covered schemas, data models, cloud tools and integration problems. Use a pipeline you worked on to check whether you can explain more than the tool names.

• Can you describe the pipeline and explain its data model?
• How would you investigate a problem with data integration?

Can you explain your BI and dashboard choices?

Power BI and data storytelling also appeared more often for this role than across all interviews. Questions covered access, refresh, performance and how people use reports.

• How would you set up dynamic row-level security in Power BI?
• Can you explain how you would handle report refresh, access controls and performance?
• Why would you choose a particular chart, and how would users explore the detail?
• How would you measure dashboard use and impact?
• Can you define a business KPI and explain uncertainty in the results?

What Data Analyst interviews cover

The 20 most asked Data Analyst questions

Counted by interview, not by mention: a question asked twice in one conversation counts once.

1. What are SQL joins, and how do INNER, LEFT, RIGHT, and FULL OUTER JOIN differ?

   Asked in 8 of 129 interviews · last asked Sep 16 · Concept · Junior

   Show answer

   SQL joins combine rows from tables using a condition. INNER JOIN returns only matching pairs of rows. LEFT JOIN keeps every left-table row, and RIGHT JOIN keeps every right-table row, adding NULLs when there is no match. FULL OUTER JOIN keeps every row from both tables, adding NULLs on the missing side.

   • A row can appear more than once if it matches several rows.
   • A WHERE filter can remove unmatched rows preserved by an outer join.
   • Support varies by database: MySQL has no native FULL OUTER JOIN.

   What the interviewer checks: The interviewer checks whether you understand which rows each join keeps and how unmatched rows are handled.

   Other ways it was asked

   • “What is a JOIN in SQL?”
   • “What is the difference between an inner join and an outer join?”
   • “What are the different types of joins in SQL?”

2. Tell me about a data quality issue you found, how you fixed it, and how you verified the fix.

   Asked in 4 of 129 interviews · last asked Oct 15 · Behavioral · Mid-level

   Show answer

   I found inflated sales in a dashboard when I compared it with the orders table. I traced the error to a join that repeated orders with multiple payment rows. I summed payments by order before joining and added a check for repeated order IDs. Before release, I matched totals to the source and checked affected orders.

   • Explain how you noticed the issue and traced it to its cause.
   • Fix the query or data process, not just the dashboard output.
   • Check corrected results against source data and add a test to catch the issue again.

   What the interviewer checks: The interviewer is checking whether you can find a data error, fix its cause, and prove the results are reliable.

   Other ways it was asked

   • “Walk me through one specific data quality issue you personally uncovered—how you spotted it, the root cause, and what you changed to resolve it.”
   • “Walk me through a specific data quality issue you discovered.”
   • “Walk me through one specific data quality issue you personally found end to end: how you detected it, confirmed the root cause, and what the permanent fix was.”

3. How would you set up dynamic row-level security in Power BI?

   Asked in 4 of 129 interviews · last asked Sep 15 · Tooling · Mid-level

   Show answer

   I would use a backend access table that maps each user's sign-in name to allowed regions or departments. In Power BI, I would connect it to the data and create an RLS role using USERPRINCIPALNAME(). After publishing, I would assign users or groups to the role and test allowed and blocked rows for different users.

   • Static RLS uses fixed rules; dynamic RLS uses the signed-in user's identity.
   • Ensure relationships carry access-table filters to the data; configure security filter direction if needed.
   • Users without a matching access entry should see no protected rows.
   • RLS applies to workspace Viewers, not Admins, Members, or Contributors.

   CodeCopy

   Role filter on the UserAccess table:
   'UserAccess'[UserUPN] = USERPRINCIPALNAME()

   What the interviewer checks: The interviewer checks whether you can connect user identity, access mappings, filter relationships, and service permissions securely.

   Other ways it was asked

   • “How do static and dynamic row-level security work in Power BI?”
   • “How would you implement dynamic row-level security in Power BI using a backend mapping table?”
   • “How do you implement row-level security in Power BI?”

4. What's the difference between WHERE and HAVING?

   Asked in 3 of 129 interviews · last asked Oct 1 · Concept · Junior

   Show answer

   WHERE filters individual rows before grouping. HAVING filters groups after aggregation, so it can test results like COUNT(*) or SUM(sales). Use WHERE for row conditions and HAVING for conditions on grouped results.

   • WHERE removes rows before aggregate values are calculated.
   • HAVING can filter aggregate results; WHERE cannot directly use aggregate functions.
   • HAVING can work without GROUP BY, treating the input as one group.

   SQLCopy

   SELECT department, COUNT(*) AS employee_count
   FROM employees
   WHERE salary > 50000
   GROUP BY department
   HAVING COUNT(*) >= 5;

   What the interviewer checks: The interviewer checks whether you understand row filtering, group filtering, and how they affect aggregate results.

   Other ways it was asked

   • “What is the difference between WHERE and HAVING in SQL?”
   • “What is the difference between the HAVING clause and the WHERE clause?”

5. What is a CTE, how does it differ from a subquery, and when would you use one?

   Asked in 3 of 129 interviews · last asked Sep 23 · Concept · Junior

   Show answer

   A CTE is a named query defined with WITH, available within one SQL statement. A subquery is a query nested inside another query. I use CTEs to split complex SQL into clear steps, reuse query logic, or handle recursion where supported. A CTE is not automatically faster than a subquery.

   • A CTE is not a stored table; its name exists only within that statement.
   • You can reference the same CTE multiple times in a statement.
   • CTEs often improve readability; subqueries work well for simple, one-time logic.
   • Performance depends on the database and query plan; a CTE does not guarantee results are calculated once.

   What the interviewer checks: The interviewer is checking whether you understand CTE scope and can choose a clear query structure without assuming performance benefits.

   Other ways it was asked

   • “What is a common table expression (CTE), and how is it used in a SQL query?”
   • “What is a common table expression in SQL?”
   • “What is the difference between a subquery and a CTE?”

6. What is the difference between RANK and DENSE_RANK?

   Asked in 3 of 129 interviews · last asked Sep 21 · Concept · Junior

   Show answer

   RANK and DENSE_RANK both give tied values the same rank. After a tie, RANK leaves gaps in the rank numbers, while DENSE_RANK uses the next rank without gaps. Both use ORDER BY inside OVER to set the ranking order.

   • RANK counts preceding rows, so ties create gaps in later ranks.
   • DENSE_RANK counts distinct ordering values, so its ranks have no gaps.
   • Use ORDER BY score DESC to give the highest score rank one.
   • Use PARTITION BY to restart ranking for each group.

   CodeCopy

   Score | RANK | DENSE_RANK
   100   | 1    | 1
   90    | 2    | 2
   90    | 2    | 2
   80    | 4    | 3

   What the interviewer checks: The interviewer is checking whether you understand ties and can choose the right ranking function.

   Other ways it was asked

   • “How do you use RANK and DENSE_RANK to rank scores in SQL?”

7. How would you identify duplicate records in a SQL table?

   Asked in 3 of 129 interviews · last asked Oct 17 · Tooling · Junior

   Show answer

   I would group by the fields that define a duplicate and use HAVING COUNT(*) > 1 to find repeated values. The fields depend on the business rules. For example, this query finds employee email addresses used in more than one row, excluding missing emails.

   • Define whether duplicates mean identical rows or matching fields, such as email.
   • Exclude unique row IDs when checking for duplicate business records.
   • GROUP BY returns duplicate groups and counts, not the full original rows.
   • Decide whether NULL values should count as duplicates.

   SQLCopy

   SELECT email, COUNT(*) AS record_count
   FROM employees
   WHERE email IS NOT NULL
   GROUP BY email
   HAVING COUNT(*) > 1;

   What the interviewer checks: The interviewer checks whether you can define duplicates correctly and use grouping to find them.

   Other ways it was asked

   • “How can you use SQL to identify duplicate records in a table?”
   • “If a company has employee records with duplicate entries, how would you identify the duplicates using SQL?”

8. How do VLOOKUP, XLOOKUP, LOOKUP, and INDEX-MATCH differ, and when would you use each?

   Asked in 3 of 129 interviews · last asked Oct 17 · Tooling · Junior

   Show answer

   VLOOKUP searches the first column and returns a value to its right; I use it for simple tables in older Excel. XLOOKUP searches either direction and defaults to exact matches; I prefer it when available. I use LOOKUP for approximate matches on ascending sorted data. INDEX-MATCH supports flexible lookups in older Excel, including looking left.

   • VLOOKUP defaults to approximate matching; use FALSE for exact matches.
   • Approximate VLOOKUP needs its first column sorted ascending; LOOKUP also needs its lookup values sorted ascending.
   • INDEX returns a value at a position; MATCH finds that position. Use MATCH with 0 for exact matching.
   • XLOOKUP is available in Microsoft 365 and Excel 2021 onward, but not Excel 2016 or 2019.

   What the interviewer checks: The interviewer checks whether you can choose a suitable lookup and avoid wrong results from matching or sorting assumptions.

   Other ways it was asked

   • “What is VLOOKUP, and how does it work?”
   • “What is the LOOKUP function, and how does it differ from INDEX and MATCH?”
   • “What is the difference between VLOOKUP, XLOOKUP, and INDEX-MATCH?”

9. What is the difference between a list and a tuple in Python?

   Asked in 3 of 129 interviews · last asked Oct 18 · Concept · Junior

   Show answer

   A list can be changed after you create it; a tuple cannot. Both keep items in order and allow duplicates. I use lists for collections I need to update, and tuples for groups of values that should stay fixed.

   • Lists use square brackets; tuples usually use parentheses, but the comma creates a non-empty tuple.
   • You can add, remove, or replace list items, but not tuple items.
   • A mutable object inside a tuple, such as a list, can still change.
   • A tuple can be a dictionary key only if all its items are hashable; a list cannot.

   What the interviewer checks: The interviewer is checking whether you understand mutability and can choose the right type.

   Other ways it was asked

   • “What is the difference between a list and a tuple?”

10. On a large unfamiliar business dataset, what specific indicators or tests do you use to assess completeness and accuracy before trusting it?

    Asked in 3 of 129 interviews · last asked Oct 15 · Scenario · Mid-level

    Show answer

    I check missing values in required fields, expected row counts, and coverage by date and business group. I test key uniqueness, valid ranges, business rules, and links between tables. I compare totals and sample records with trusted sources. I investigate unusual values and confirm expected exceptions with the data owner before using the data.

    • Measure missing-value rates by field and group; check for missing records, dates, and groups too.
    • Check duplicates against what each row should represent, and flag unmatched links between tables.
    • Test formats, ranges, and business rules; unusual values need investigation, not automatic removal.
    • Passing validation does not prove accuracy; compare records and totals with reliable sources.

    What the interviewer checks: The interviewer checks whether you can test both completeness and accuracy using business expectations and trusted sources.

    Other ways it was asked

    • “When you receive a large, unfamiliar business dataset, what specific checks do you run to assess accuracy and completeness before you start analyzing it?”
    • “When you receive a large unfamiliar business data set, what specific checks do you run to assess accuracy and completeness before doing any analysis?”

11. When you’re reviewing a high-volume batch of business documents, what criteria do you use to decide which documents get a full review versus a spot check?

    Asked in 3 of 129 interviews · last asked Oct 15 · Scenario · Mid-level

    Show answer

    I fully review documents with high financial or legal risk, unusual details, known errors, or rules requiring a full check. I spot-check routine, lower-risk documents from sources with a reliable history, using a random sample. If the sample shows problems, I expand the review.

    • Follow company policies and any rules requiring full review.
    • Consider financial impact, sensitive data, unusual details, and past error rates.
    • Sample across document types and sources, not just the easiest items.
    • Set clear rules for expanding checks when errors appear.

    What the interviewer checks: The interviewer is checking whether you can manage a large workload without missing important risks.

    Other ways it was asked

    • “When you're reviewing high volumes of business documents, what criteria do you use to decide which items must be reviewed in full versus eligible for spot checking?”
    • “When reviewing high volumes of business documents, what criteria do you use to decide which items get full review versus spot checks?”

12. How would privacy regulations or frameworks such as GDPR, CCPA/CPRA, HIPAA, SOC 2, or ISO 27001 influence the specific data controls you implement?

    Asked in 3 of 129 interviews · last asked Oct 15 · Scenario · Senior

    Show answer

    I would turn the requirements that apply to our data into specific, tested controls. These could include role-based access, masked reports, encryption, audit logs, and retention limits. I would confirm the scope with legal and security teams, document the controls, and check that they work.

    • GDPR: confirm a lawful basis, collect only needed data, and support applicable access and deletion rights.
    • CCPA/CPRA: support applicable consumer rights, including opting out of personal data sales or sharing.
    • HIPAA: protect covered health information through access controls, audit logs, and limits on use and disclosure.
    • SOC 2 assesses controls; ISO 27001 sets security management requirements. Both need evidence, not just written policies.

    What the interviewer checks: The interviewer checks whether you can connect applicable requirements to practical controls without treating every rule as identical.

    Other ways it was asked

    • “How have privacy regulations or frameworks such as GDPR, CCPA/CPRA, ISO 27001, or SOC 2 shaped your data-handling practices?”
    • “How have privacy regulations or frameworks such as GDPR, HIPAA, PCI DSS, SOC 2, or ISO 27001 changed the way you handle business data?”
    • “How have privacy regulations or frameworks such as GDPR, CCPA/CPRA, HIPAA, SOC 2, or ISO 27001 influenced the specific data controls you implemented?”

13. Can you share one concrete example where business context, not just a tech validation rule, helped you catch a document or data issue that automated checks would likely miss?

    Asked in 3 of 129 interviews · last asked Oct 15 · Behavioral · Mid-level

    Show answer

    Hypothetical example: I caught a repeated setup fee because the supplier’s contract allowed it only once. Both invoices had valid dates, different invoice numbers, and correct totals, so basic checks passed. I checked the contract and earlier payment, confirmed the duplicate charge with accounts payable, and stopped the second payment.

    • Valid fields and correct totals do not prove a charge is justified.
    • The contract provided the business context that basic invoice checks lacked.
    • Checking the earlier payment confirmed the issue before taking action.

    What the interviewer checks: The interviewer is checking whether you use business knowledge to spot issues, verify them, and prevent harm.

    Other ways it was asked

    • “Can you give a concrete example where business context, rather than a technical validation rule, helped you catch an error in a document or record that otherwise looked valid?”
    • “Can you share a specific example where business context helped you catch a problem that basic technical validation like formats, totals, or required fields would likely miss?”

14. When reviewing an invoice or contract, what specific signals make you suspect the content is inaccurate, incomplete, or internally inconsistent?

    Asked in 3 of 129 interviews · last asked Oct 15 · Scenario · Mid-level

    Show answer

    I look for totals that do not add up, conflicting dates or terms, and charges that do not match agreed prices or delivered work. I also check for duplicate invoices, unexplained bank detail changes, and unclear obligations. I compare concerns with purchase orders, delivery records, and signed agreements before calling them errors.

    • Quantities, rates, discounts, and taxes should explain the total; allow for valid rounding differences.
    • Names, dates, payment terms, and amounts should agree across sections and supporting records.
    • Duplicate charges, unexpected fees, or changed bank details need checking, but do not prove an error.
    • Unclear scope, missing attachments, or conflicting clauses can leave contract obligations incomplete or inconsistent.

    What the interviewer checks: The interviewer is checking whether you spot business inconsistencies and verify them with evidence, rather than only checking missing fields.

    Other ways it was asked

    • “When reviewing an invoice or contract, what are the top signals that tell you it's internally inconsistent or likely inaccurate beyond simple missing fields?”
    • “When reviewing an invoice or contract, what specific business red flags tell you the content is inaccurate, incomplete, or internally inconsistent?”

15. How would you clean a dataset and address common data quality problems?

    Asked in 3 of 129 interviews · last asked Sep 16 · Scenario · Junior

    Show answer

    I first check what each field means, then look for missing values, duplicates, wrong types, and inconsistent formats. I choose fixes based on business rules, such as correcting dates or removing confirmed duplicates. I investigate unusual values before changing them. I keep the raw data, record my changes, and check the results.

    • Check row counts, missing values, value ranges, and whether keys should be unique.
    • Handle missing values based on their cause and use; do not fill or drop them blindly.
    • Standardize types, labels, dates, and units; confirm duplicates before removing them.
    • Check unusual values with the source, then validate cleaned data against business rules.

    What the interviewer checks: The interviewer checks whether you can find quality problems and fix them without losing valid information.

    Other ways it was asked

    • “What data-quality problems have you addressed, and what approaches did you use?”
    • “How would you clean a dataset?”
    • “When you receive a new dataset, what steps do you take to clean and prepare it?”

16. How do you set up a data gateway in Power BI?

    Asked in 3 of 129 interviews · last asked Sep 15 · Tooling · Mid-level

    Show answer

    A gateway lets Power BI Service securely access on-premises data. I install the standard gateway on an always-on Windows machine that can reach the data source. I sign in, register it, and save the recovery key. In Power BI Service, I add a connection, enter credentials, link the semantic model, and test refresh.

    • Use standard mode for shared team access; personal mode supports only Power BI and has fewer features.
    • Install required data source drivers and allow the gateway’s outbound network connections.
    • Match connection details, such as server and database, and give users permission to use the connection.
    • Schedule refresh for imported data; DirectQuery uses the gateway when querying on-premises sources.

    What the interviewer checks: The interviewer checks whether you can connect Power BI to local data securely and keep access reliable.

    Other ways it was asked

    • “How do data gateways work in Power BI, and how do you configure them?”
    • “How do you set up an on-premises data gateway in Power BI?”

17. How would you approach analyzing a large dataset, from sourcing data to choosing processing methods?

    Asked in 2 of 129 interviews · last asked Sep 11 · Scenario · Mid-level

    Show answer

    I start with the business question and find reliable data sources I am allowed to use. I check data quality and keep only needed rows and columns. I choose SQL, Python, or distributed tools based on data size, location, memory, and speed needs. I validate the results and explain the findings and limits.

    • Confirm what each source covers, how fields are defined, and whether access is approved.
    • Check missing values, duplicates, and join keys before analysis.
    • Filter and summarize in the database; consider chunks before moving to distributed processing.
    • Check results against known totals and explain any sampling limits.

    What the interviewer checks: The interviewer checks whether you can plan reliable analysis and choose tools that fit the data and available resources.

    Other ways it was asked

    • “How do you work with a large dataset, and why would you choose that approach?”
    • “Can you tell me about a time when you analyzed a large dataset and walk me through the steps, including where you got the data?”

18. What is data?

    Asked in 2 of 129 interviews · last asked Oct 9 · Concept · Junior

    Show answer

    Data is a collection of recorded facts, measurements, or observations. It can include numbers, text, images, or sounds. Analysts use data to answer questions and support decisions.

    • Data can be numerical or non-numerical.
    • Data needs context to be understood correctly.
    • Data may contain errors or missing values, so its quality must be checked.

    What the interviewer checks: The interviewer is checking whether you can explain a basic concept clearly and connect it to analysis.

19. What does a data analyst do, and how is it different from a data scientist?

    Asked in 2 of 129 interviews · last asked Sep 8 · Concept · Junior

    Show answer

    A data analyst cleans and studies data to answer business questions and help people make decisions. They often use SQL, spreadsheets, charts, and dashboards to explain results. A data scientist often builds statistical or machine learning models to make predictions. The roles overlap, and duties vary by company.

    • Analysts turn business questions into data analysis and explain what the results mean.
    • Data scientists often focus more on predictive models and machine learning.
    • Both roles need clean data, statistical skills, and clear communication.
    • Job titles do not always mean the same duties across companies.

    What the interviewer checks: The interviewer wants to know whether you understand the analyst’s business purpose and how the two roles overlap.

    Other ways it was asked

    • “What does a data analyst do?”

20. What are SQL window functions, and when would you use one?

    Asked in 2 of 129 interviews · last asked Oct 1 · Concept · Mid-level

    Show answer

    SQL window functions calculate values across related rows without combining those rows into one result row. They use an OVER clause to define which rows to use and how to order them. I use them for rankings, running totals, or comparisons with a previous row.

    • Unlike GROUP BY, window functions keep individual rows in the result.
    • PARTITION BY separates rows into groups; ORDER BY sets the order within each group.
    • Common examples include ROW_NUMBER, RANK, LAG, and SUM used with OVER.
    • For running totals, an explicit window frame controls which rows contribute.

    SQLCopy

    SELECT
      employee_id,
      department,
      salary,
      AVG(salary) OVER (PARTITION BY department) AS department_avg
    FROM employees;

    What the interviewer checks: The interviewer checks whether you can calculate across related rows while keeping each row visible.

    Other ways it was asked

    • “What are SQL window functions, and what are some common examples?”
    • “What is a SQL window function?”

Related interview questions

• SQL interview questions
• ETL and Data Pipelines interview questions
• Power BI interview questions
• Data Engineer interview questions
• Data Scientist interview questions
• Machine Learning Engineer interview questions
• AI Engineer interview questions
• AI Trainer interview questions

How this list is made

Every question here was asked by a real interviewer in a real job interview that a candidate ran MeetAssist in. After the call, the interviewer's questions are pulled from the transcript. Practice sessions, tests, exams and anything that was not a job interview are dropped.

Questions are rewritten so they make sense on their own: names of people, employers and clients are removed, and speech-to-text mistakes are fixed. Small talk, logistics and questions about one candidate's own résumé are left out. Rephrasings of the same question are merged, and the count is the number of different interviews it came up in.

The interviews come from candidates who use MeetAssist, so they lean towards software, data and AI roles: the shares describe these interviews, not every Data Analyst interview. Answers to the most asked questions are written and checked by an editor; a "common mistake" is shown only when reviews of real answers in these interviews show it, with the count.

129 interviews, 75 candidates, Nov 2025 to Oct 2026. Updated every month with new interviews.