Power BI Interview Preparation Plan

If you are applying for data analyst roles at Accenture, Infosys, TCS, Cognizant, Deloitte, or a mid-size analytics firm in Bengaluru, Hyderabad, Pune, or Gurugram, Power BI now shows up in almost every interview. The good news: Power BI interviews are predictable. The questions cluster around five areas, and a focused two-week plan will put you ahead of most freshers who only "watched tutorials".

This is that plan - a topic checklist, the questions you will actually be asked with model answers, how to talk through a project end-to-end, and a day-by-day schedule.

The complete topic checklist

Interviewers draw from the same five buckets. Treat this as a tick-list: if you can teach each line to a friend, you are ready.

Power Query (data prep, the M layer)

• Remove duplicates, change data types, split and merge columns.
• Unpivot wide data into a tall, model-friendly shape.
• Merge (join) vs. Append (stack) queries - and when to use each.
• Query folding: pushing transformations back to the source so the database does the work.
• Parameters and a basic "source path" parameter for portability.

Data modeling and star schema

• Fact vs. dimension tables; why Power BI prefers a flat star schema over a normalized snowflake.
• Cardinality: one-to-many, one-to-one, many-to-many.
• Single vs. both cross-filter direction, and why "both" is risky.
• Active vs. inactive relationships and USERELATIONSHIP.
• A dedicated, marked Date dimension - the foundation for all time intelligence.

DAX (the scoring round)

• Measure vs. calculated column (query time vs. refresh time).
• Row context vs. filter context - the single most-tested idea.
• CALCULATE as the function that modifies filter context.
• Filter modifiers: ALL, ALLEXCEPT, REMOVEFILTERS, KEEPFILTERS.
• Iterators: SUMX, AVERAGEX, RANKX.
• Time intelligence: TOTALYTD, SAMEPERIODLASTYEAR, DATEADD.

Visuals and UX

• Choosing the right visual; when NOT to use a pie chart (more than 3-4 categories).
• Slicers, drill-through, drill-down, and bookmarks for navigation.
• Conditional formatting and data bars to draw the eye.
• Tooltips, report-level vs. page-level vs. visual-level filters.

Power BI Service, RLS, and refresh

• Workspaces, publishing, and apps.
• Scheduled refresh and the on-premises data gateway.
• Incremental refresh for large tables.
• Row-Level Security (RLS): static roles and dynamic RLS with USERPRINCIPALNAME().

Build one small model that touches all five areas. Use a kirana-store or Swiggy-style dataset: a Sales fact table (date, product, quantity, amount in ₹) with Product, Store, and Date dimensions. Connecting these correctly teaches star schema faster than any video.

Interview questions with model answers

Practice saying these out loud. Fluency under pressure is the real skill - interviewers can tell within two answers whether you understand context or are reciting.

1. Why do we prefer a star schema in Power BI?
It keeps relationships simple and single-directional, makes filter context predictable, compresses better in the VertiPaq engine, and queries faster than a normalized snowflake with many joins.

2. Difference between a measure and a calculated column?
A calculated column is computed at refresh, stored in the model, and consumes memory - use it when you need a value per row to slice or relate on. A measure is computed at query time inside the current filter context and stores nothing - use it for aggregations like totals and ratios. Default to measures.

3. What is the difference between row context and filter context?
Row context is "the current row" - it exists inside calculated columns and iterators like SUMX. Filter context is the set of filters applied by slicers, rows/columns of a visual, and CALCULATE. A measure has no row context until an iterator creates one; CALCULATE is how you change filter context.

4. What does CALCULATE do?
It evaluates an expression in a modified filter context. It is the only function that can change filter context, which makes it the backbone of nearly every non-trivial measure.

5. Difference between ALL, ALLEXCEPT, and REMOVEFILTERS?
ALL removes all filters from a table or columns. ALLEXCEPT removes filters from a table except the columns you list. REMOVEFILTERS does the same job as ALL but is read-only (cannot return a table) and reads more clearly in modern DAX. A classic use is "% of total":

% of Total Sales =
DIVIDE (
    [Total Sales],
    CALCULATE ( [Total Sales], ALL ( 'Product' ) )
)

6. Write a year-to-date and a year-over-year measure.

Total Sales = SUM ( Sales[Amount] )

Sales YTD = TOTALYTD ( [Total Sales], 'Date'[Date] )

Sales LY =
CALCULATE ( [Total Sales], SAMEPERIODLASTYEAR ( 'Date'[Date] ) )

YoY % =
DIVIDE ( [Total Sales] - [Sales LY], [Sales LY] )

Mention that all of these require a continuous, marked Date table - a common follow-up trap.

7. SUM vs. SUMX - when do you need the iterator?
SUM aggregates one column. SUMX iterates row by row, evaluating an expression per row before summing - use it when the calculation does not exist as a single column:

Revenue = SUMX ( Sales, Sales[Quantity] * Sales[Unit Price] )

8. How do you handle a many-to-many relationship?
Ideally introduce a bridge dimension table so each side becomes one-to-many. Power BI also supports a native many-to-many cardinality, but it is slower and the filter behaviour is harder to reason about - mention the bridge first.

9. What is query folding and why does it matter?
When Power Query can translate your steps into a single native query (e.g. SQL) that the source runs, that is query folding. It pushes work to the database, so refreshes are far faster and lighter. Steps like custom M functions can break folding - keep heavy transforms early and source-friendly.

10. What is Row-Level Security and how do you make it dynamic?
RLS restricts the rows a user sees. A static role hard-codes a filter (e.g. Region = "South"). Dynamic RLS uses the logged-in user to filter automatically, so one role serves everyone:

[Email] = USERPRINCIPALNAME ()

You apply this as a table filter in a security role, mapping each user's email to their allowed rows via a user-mapping table.

11. How would your report perform on 1 crore rows?
Reduce cardinality, prefer measures over calculated columns, use aggregation tables for summary visuals, set up incremental refresh so only new partitions reload, and remove unused columns to shrink the model in memory.

12. Difference between Import, DirectQuery, and Live connection?
Import loads data into VertiPaq - fastest, but needs refresh. DirectQuery queries the source live - always current, but slower and DAX-limited. Live connection points to an existing Analysis Services / Power BI dataset. For most analyst roles, Import is the default answer.

How to talk through a Power BI project end-to-end

Almost every interview ends with: "Walk me through a dashboard you built." A vague answer ("I made a sales dashboard") sinks strong candidates. Tell it as a pipeline, in order, so the interviewer hears that you understand the whole flow.

1. Business question - "City managers for a cab aggregator across Mumbai, Delhi, and Bengaluru needed to see which routes were losing money and where demand peaked. About 2 lakh trip records over three years."
2. Source and ingestion - where the data came from (CSV/SQL), connected via Power Query.
3. Cleaning - "In Power Query I removed duplicate trip IDs, fixed data types, and handled missing fares by flagging rather than dropping them."
4. Modeling - "I built a star schema: a Trips fact table with Date, City, and Driver dimensions, all one-to-many, single-direction."
5. DAX - "I wrote measures for revenue per km, utilisation %, and a YoY growth measure using SAMEPERIODLASTYEAR."
6. Visualisation - "A city-level overview page with drill-through to a route detail page, conditional formatting to flag loss-making routes."
7. Insight and outcome - "It surfaced that late-night airport routes had the highest margin - the actual decision the dashboard drove."

Then be ready for probing follow-ups, which exist to confirm the project is genuinely yours: "Why that visual?", "How did you handle missing fares?", "How would this scale to 1 crore rows?" Always end on the insight, not the chart count.

Your 2-week prep schedule

You do not need eight hours a day. Ninety focused minutes works. Each day: ~30 min concept (written by hand), ~30 min building your model, ~30 min answering questions out loud and recording yourself.

Week 1 - fundamentals and modeling

• Day 1: Power Query - load the kirana dataset, remove duplicates, set types, unpivot.
• Day 2: Merge vs. append, query folding; build your Date dimension.
• Day 3: Star schema - relationships, cardinality, cross-filter direction.
• Day 4: Row vs. filter context (write it out as a one-page note).
• Day 5: CALCULATE plus ALL / ALLEXCEPT / REMOVEFILTERS.
• Day 6: Visuals and UX - slicers, drill-through, bookmarks, conditional formatting.
• Day 7: Mock round 1 - answer questions 1-6 out loud, recorded.

Week 2 - DAX depth, Service, and delivery

• Day 8: Iterators - SUMX, RANKX; SUM vs. SUMX.
• Day 9: Time intelligence - YTD, YoY, DATEADD.
• Day 10: Power BI Service - publish, scheduled refresh, gateway.
• Day 11: RLS - static and dynamic with USERPRINCIPALNAME().
• Day 12: Performance - aggregations, incremental refresh, Import vs. DirectQuery.
• Day 13: Polish the project story; rehearse the end-to-end talk-through.
• Day 14: Full mock - project walk-through plus questions 7-12.

The PL-300 angle

If you are targeting roles that value certification, the PL-300: Microsoft Power BI Data Analyst exam maps almost one-to-one with this checklist: prepare data, model data, visualize and analyze, deploy and maintain. Preparing for the interview and the exam together is efficient - the same DAX and modeling knowledge serves both.

Why this matters for your package

In 2026, freshers who genuinely know SQL plus Power BI get shortlisted noticeably faster and tend to earn ₹1-1.5 LPA more than Excel-only candidates, with entry packages commonly in the ₹4-7 LPA band and higher in Bengaluru and Hyderabad (upGrad, GrowAI). A confident Power BI round is one of the cheapest ways to move into that higher band.

Common mistakes to avoid

• Memorizing DAX without understanding filter context - interviewers will catch you.
• Building a beautiful dashboard you cannot explain.
• Ignoring Power Query - data cleaning questions are common and easy marks.
• Forgetting a proper Date table, then fumbling every time-intelligence question.
• Never doing a mock interview. Speaking the answers is a separate skill from knowing them.

FAQ

What DAX questions are asked in Power BI interviews?
Expect CALCULATE and filter context, measure vs. calculated column, SUM vs. SUMX, the ALL / ALLEXCEPT family, and time intelligence (YTD and YoY). For analyst roles you are usually asked to explain and write short snippets, not produce 50-line measures.

How do I explain a Power BI project in an interview?
Tell it as a pipeline: business question, source, Power Query cleaning, star-schema model, key DAX measures, visuals and navigation, and the final insight or decision it drove. Keep it under two minutes and finish on the business outcome, not the visual count.

Is DAX asked for fresher roles?
Yes, but at a basic level. Freshers are expected to explain measures vs. columns, filter context, and write simple measures like a total, a YTD, and a YoY %. Deep DAX optimisation is for experienced roles - solid fundamentals are enough to clear a fresher round.

Power BI or SQL - which matters more for an analyst interview?
Both, and they are tested separately. SQL screens you in (joins, group by, window functions); Power BI proves you can model and present. Learn SQL first, then Power BI - the combination is what moves you into the higher salary band.

How long does it take to prepare for a Power BI interview?
With a real project already built, this two-week, 90-minutes-a-day plan is enough to be interview-ready. From zero, give yourself four to six weeks so you have time to build and genuinely understand one defensible project.

Do I need the PL-300 certification to get hired?
No - a project you can defend usually carries more weight than a certificate. PL-300 helps you clear resume screens and structures your learning, so treat it as a useful bonus, not a prerequisite.

Related: Power BI Portfolio Projects That Get You Hired · Build your skill score

Don't just read. Prove your skill on DevWithData.