Interview Questions for Power BI Data Analysts
· 7 min read
Most Power BI interview questions for data analyst roles test four things: whether you can model data as a star schema, whether you understand the difference between a measure and a calculated column, whether you can explain filter context and CALCULATE, and whether you can debug a number that looks wrong. If you can answer the questions below in your own words and write the DAX without looking, you are ready for a junior or mid-level interview.
-- The measure you will be asked to write in almost every interview
% of Total Revenue =
DIVIDE ( [Revenue], CALCULATE ( [Revenue], REMOVEFILTERS ( Products[Category] ) ) )The examples use a Brightline Supply Co. model: a Sales fact table (one row per order line) related to Products, Customers and Date dimensions.
Modeling questions
1. What is a star schema and why does Power BI prefer it?
A star schema has one fact table of events (sales lines) in the middle and dimension tables (products, customers, dates) around it, each joined one-to-many to the fact. Filters flow from the dimensions to the fact. It keeps DAX simple, makes slicers behave predictably and compresses well. A single wide flat table works for small reports but gets slow and confusing as it grows.
2. What do cardinality and cross-filter direction mean?
Cardinality is how rows match: one-to-many (one product, many sales lines) is the normal case. Cross-filter direction says which way filters flow. Single direction (dimension filters fact) is the default and the safe choice. Both directions can be needed for some many-to-many cases, but it can create ambiguous paths and slower reports, so use it deliberately.
3. Import vs DirectQuery?
Import loads a compressed copy of the data: fast, full DAX support, as fresh as the last refresh. DirectQuery queries the source at view time: current, but slower and with some limits. Composite models mix both.
4. Why do you need a date table?
Time intelligence (YTD, prior year) needs a continuous calendar with one row per day. You create it with CALENDAR or in Power Query, relate it to the fact table's date, and mark it as a date table. Full walkthrough: how to create a date table.
DAX interview questions
5. Measure vs calculated column?
A calculated column is computed per row at refresh and stored. A measure is computed at query time in the filter context of each visual cell. Use columns for things you slice or group by, and measures for anything you aggregate, especially ratios. Details and examples: measure vs calculated column.
6. What is filter context vs row context?
Filter context is the set of filters on the model for the current evaluation: visual rows, slicers, page filters, CALCULATE filters. Row context exists while DAX iterates a table (calculated columns, SUMX, FILTER) and means "the current row". Row context does not filter anything on its own.
7. What does CALCULATE do? What is context transition?
CALCULATE evaluates an expression in a modified filter context: it adds, replaces or removes filters. When CALCULATE runs inside a row context, it turns the current row into a filter. That is context transition, and it is why a measure called inside SUMX returns the value for the current row. See DAX CALCULATE explained.
8. SUM vs SUMX?
SUM adds one column. SUMX evaluates an expression per row, then sums. On order lines with quantity and price, SUMX ( Sales, Sales[Quantity] * Sales[UnitPrice] ) is revenue. SUM ( Sales[Quantity] ) * SUM ( Sales[UnitPrice] ) is a meaningless number. See SUM vs SUMX.
9. ALL vs ALLSELECTED vs REMOVEFILTERS?
ALL and REMOVEFILTERS both remove filters when used inside CALCULATE (REMOVEFILTERS can only do that, so it reads clearer). ALL can also be used as a table function, returning all rows. ALLSELECTED removes filters from the visual but keeps slicer and page selections, so percentages add up to 100% of what the user selected.
10. How do you use an inactive relationship?
CALCULATE ( [Revenue], USERELATIONSHIP ( Sales[ShipDate], 'Date'[Date] ) ). USERELATIONSHIP only works inside CALCULATE (or another function's filter argument), and the relationship must already exist in the model. See RELATED and USERELATIONSHIP.
11. DAX vs Power Query: where should a calculation go?
Row-level cleaning and new columns that do not depend on report filters belong in Power Query, during load. Aggregations that must respond to filters belong in DAX measures. Covered in DAX vs Power Query.
Power BI scenario based interview questions
Interviewers like scenarios because they show whether you can debug, not just define. A good answer names the likely cause, how you would check it, and the fix.
12. "The total row does not equal the sum of the rows. Is it a bug?"
Usually not. A measure is evaluated on the total row with no row filter, so it computes the total from scratch. That is correct for ratios (total margin % is not the sum of category margins) and for distinct counts (a customer who bought in two regions counts once in the total). It is wrong only when the business wants a sum of row values, in which case you iterate: SUMX ( VALUES ( Customers[Region] ), [Measure] ).
13. "Your YTD measure shows the same value as the monthly number."
Check that the measure uses the date table's column ('Date'[Date]), not the fact table's date, that the visual's month comes from the date table, and that the date table is related and marked. A fact-table date column has gaps and does not clear the month filter.
14. "A slicer on Region does not filter one of the visuals."
Check the relationship path from Customers to that visual's fact table (missing, inactive, or filtering in the wrong direction), then check Edit interactions on the slicer. If the visual uses a different fact table with no customer key, the slicer cannot filter it without a model change.
15. "The business says revenue is too high."
Compare the measure to an independent source (a SQL query or the finance figure). Common causes: cancelled and refunded orders not excluded, discount not applied, a many-to-many or bi-directional relationship duplicating rows, or SUM of a pre-calculated column that was already a total. On Brightline's sample, revenue including every status is 744.10, while completed-only revenue is 612.10. Knowing the business definition is half the answer.
16. "An AI assistant wrote this measure for % of total. What is wrong?"
% of Total =
DIVIDE ( [Revenue], CALCULATE ( [Revenue], ALL ( Products[ProductName] ) ) )This is the kind of question that now appears in interviews because analysts use AI tools daily. With Category on the rows, every row shows 100%: ALL removes the filter on ProductName, but the visual filters Category, so numerator and denominator are the same. The fix is REMOVEFILTERS ( Products[Category] ), or REMOVEFILTERS ( Products ) if the visual can be grouped by any product column. The skill being tested: do not trust a formula because it runs. Check one row by hand.
17. "How would you restrict each regional manager to their own region?"
Row-level security. Create a role with a DAX filter on Customers, such as [Region] = "West" for a static role, or a dynamic rule that compares a manager table's email column with USERPRINCIPALNAME (). Assign users to roles in the Power BI service and test with "View as".
How to prepare
- Build a small star schema yourself (a fact table and three dimensions) and write each measure above from memory.
- For every measure, compute one cell by hand and compare. Interviewers love "how do you know it is right?"
- Practise explaining filter context out loud with a concrete visual: "in the Cleaning row, the filter is Category = Cleaning, so..."
- Work through DAX practice exercises with known answers.
- Brush up on SQL too. Many analyst interviews combine both: see SQL interview questions for data analysts and DAX vs SQL.
Common mistakes in interviews
- Defining without an example. "A measure is dynamic" is weak. "Margin % must be a measure because each category needs its own division" is strong.
- Reaching for bi-directional filters to fix a visual. Interviewers see this as a red flag. Explain the relationship path instead.
- Trusting generated DAX. If you use an AI assistant in a take-home task, say how you verified its output. Wrong filter context (ALL on the wrong column, FILTER over the whole fact table) and SUM where SUMX was needed are the mistakes these tools make most.
FAQ
What Power BI questions are asked in a data analyst interview?
Modeling (star schema, relationships, date table), DAX (measure vs column, CALCULATE, filter context, SUMX, time intelligence), Power Query basics, and one or two scenario questions about a number that looks wrong.
What are the most common DAX interview questions?
Measure vs calculated column, filter vs row context, what CALCULATE does, SUM vs SUMX, ALL vs ALLSELECTED, and writing a YTD or % of total measure.
How do I answer scenario based Power BI questions?
Name the likely cause, say how you would confirm it (check a cell by hand, inspect the relationship, compare with a SQL query), then give the fix. Interviewers care about the debugging process as much as the answer.
Do I need to know Power Query for a Power BI interview?
The basics, yes: importing, changing types, merging and appending queries, unpivoting and adding conditional columns. Junior interviews usually focus more on DAX and modeling.
Practise this on a realistic dataset
Solve it yourself on one company's data, with instant feedback that points at what's wrong without giving away the answer.
- Audit: every region sold the same?Audit
- Audit: a year to date that never growsAudit
- Audit: every segment at 100%Audit