A sales manager asks for a pipeline dashboard for 18 representatives by Friday but only says it should show what needs attention. What would you clarify first?

I would first identify the decision the manager must make from the dashboard. - Ask whether the Friday meeting reallocates leads, reviews quota risk, or follows overdue deals, because each decision needs a different primary KPI. - Define the audience and cadence as 1 manager plus 18 representatives using a daily snapshot, then record which owner, stage, and region fields they need to filter. - Leave the meeting with a 1-page wireframe and 3 signed examples of deals that should appear as needing attention, then use those rows as acceptance checks.

A finance dashboard has 12,000 invoice rows, including cancellations and partial refunds, and the CFO asks for one monthly revenue KPI. How would you define it before building the card?

I would write a metric contract for monthly net revenue before creating the KPI card. - Set the formula to paid invoice amount minus posted refunds, exclude canceled invoices, and state whether the month comes from payment date or invoice date. - Fix the display grain at 1 calendar month and document currency conversion, tax treatment, and the cutoff time for the daily refresh. - Reconcile 3 sample months to the finance ledger and require a zero unexplained difference before the card is accepted.

A product funnel dashboard uses 800,000 event rows, and one user can trigger checkout_started five times. What grain would you choose for a daily conversion KPI?

I would define the KPI at one user per funnel step per calendar day, not one raw event row. - Deduplicate the 800,000 events by user_id, event_date, and step so 5 checkout attempts by 1 user count once for that step. - Calculate daily conversion as distinct purchasers divided by distinct eligible funnel entrants, with both populations using the same cohort rule. - Validate the grain with a 20-row user journey sample and add a grain note beside the dataset definition.

You have one page for 6 executives who review revenue, margin, and target attainment in 5 minutes each Monday. How would you choose the dashboard layout and charts?

I would build a summary page that shows status first and supporting trends second. - Put 3 KPI cards for revenue, margin, and target attainment in the first row, each with the current value and variance to target. - Use 1 monthly line chart for the 12-month trend and 1 sorted bar chart for the 6 regions, because executives need direction and comparison rather than row-level detail. - Keep the page to about 5 visuals, place the latest refresh timestamp visibly, and test that all 3 questions can be answered within a 5-minute walkthrough.

A marketing lead wants to compare weekly qualified leads for the last 26 weeks against a target of 400. Which visual would you use and why?

I would use a line chart with a separate target reference line. - Put week on a continuous x-axis and qualified lead count on the y-axis so the 26-week direction and seasonality remain visible. - Add a clearly labeled horizontal line at 400 rather than a second bar series, which makes misses and recoveries easy to scan. - Check the chart at 1366 pixels wide and verify that all 26 points are readable without labels on every mark.

A support dashboard must compare 14 issue categories by ticket count for 25 team leads. The current pie chart has unreadable slices. What would you replace it with?

I would replace the pie with a descending horizontal bar chart. - Sort all 14 categories by ticket count so the largest workload is visible without comparing angles or hunting through a legend. - Show direct count labels and optionally group categories below 1 percent into Other only if the category detail remains available elsewhere. - Test the visual with the longest category name and a 1280-pixel screen to ensure labels do not truncate or overlap.

An operations director needs to see whether yesterday's 92 percent on-time delivery KPI met a 95 percent target and how it changed from the prior day. How would you present it?

I would use one KPI card with the actual, target variance, and prior-day change shown together. - Display 92 percent as the primary value and a labeled minus 3 percentage-point variance to the 95 percent target. - Add the prior-day comparison as a secondary value, such as down 1.2 percentage points, without mixing percentage points with percent change. - Use text or an icon in addition to color and verify the card remains understandable in grayscale and at 200 percent zoom.

Thirty account managers need a daily dashboard to find customers with renewals due in the next 14 days, including owner, value, and last contact. Would you lead with a chart or a table?

I would lead with an actionable table because users must identify specific customer rows. - Include customer, renewal date, owner, contract value, days remaining, and last-contact date, then sort by days remaining ascending. - Add filters for the 30 owners and renewal-value band, while keeping a small card above the table for total value due within 14 days. - Validate the table with 10 known accounts and check that the first screen shows enough columns without horizontal scrolling at the standard laptop width.

Sales defines an active customer as anyone with an order in 90 days, while Customer Success uses a login in 30 days. Both teams want one active customers KPI on a shared dashboard in 4 days. What would you do?

I would not merge the two definitions into one unlabeled KPI. - Document 2 candidate definitions with formulas, date fields, and example customer IDs, then ask the dashboard owner to approve the business meaning. - If both decisions matter, show 2 explicitly named measures such as Purchasing active, 90 days and Product active, 30 days rather than hiding the difference. - Test 10 boundary customers against both rules and attach the approved metric contract before the 4-day build starts.

You have 3 days to deliver a first inventory dashboard from 2 million stock-movement rows for 8 warehouse supervisors. How would you define a junior-sized first version?

I would scope the first version to the supervisors' daily stock-risk decision and make its acceptance criteria explicit. - Limit version 1 to 3 KPIs, current stock, items below reorder point, and stock value, plus 1 warehouse filter and 1 item-level table. - Write a wireframe and a field mapping from each visual to the source column or measure before authoring the dashboard. - Agree on 5 acceptance checks, including known low-stock SKUs from at least 2 warehouses, and complete a review with 2 supervisors before the 3-day deadline.

A Power BI report has 4 pages and 12 visuals, but the regional manager wants West applied only to one chart, 2026 applied to one page, and Active applied everywhere. Where would you place each filter?

I would use visual-level, page-level, and report-level filters for those three scopes. - Put Region equals West on the single chart's visual filter so the other 11 visuals keep their own region context. - Put Year equals 2026 on the relevant page filter so every visual on that page receives it. - Put Status equals Active on the report-level filter, then inspect 1 visual on each of the 4 pages to confirm the filter scope.

A Power BI page serves 40 sales users who need to switch among 5 regions and 3 product families while keeping the selected year visible. How would you configure the slicers?

I would add clearly labeled slicers bound to dimension fields and keep their state easy to inspect. - Use a single-select Year slicer when the KPI specification allows only 1 year, and multi-select slicers for the 5 regions and 3 product families. - Place the slicers consistently at the top of the report canvas and show the current selection in a subtitle or filter summary. - Test 3 combinations against a simple table of expected totals and use Edit interactions if a visual must intentionally ignore one slicer.

A Tableau dashboard has a bar chart for 8 regions and a detail table with 60,000 order rows. Users want a click on a region to narrow the table. How would you build it?

I would create a Tableau filter action from the region chart to the detail table. - Set the source sheet to the 8-region bar chart, run the action on Select, and target only the order-detail sheet. - Map Region to Region and choose the clearing behavior deliberately, such as Show all values when the selection is cleared. - Click 3 regions and compare the target row counts with worksheet filters to confirm that the action passes the intended field.

In Looker, 12 product managers need a weekly table of orders and revenue by category from an existing Explore. What authoring steps would you take?

I would build the query from governed fields in the existing Explore, verify it, and save it as a dashboard tile. - Select Week and Product Category as dimensions plus Orders and Revenue as measures, then filter to the required date range. - Run the query and check 3 category totals against known source results before changing visualization settings. - Save the result to the shared dashboard with a clear title and preserve the field definitions from LookML rather than recreating KPI logic in the tile.

A Power BI summary page shows sales for 20 regions, and users must open a page listing the orders behind the selected region. How would you configure drill-through?

I would create a dedicated order-detail page and add Region to its drill-through filter well. - Put an order table on the target page with order ID, date, customer, and amount, and add a Back button for the return path. - Users can right-click a regional data point and choose the drill-through page, which carries that region's filter context. - Test 2 regions and confirm both the detail row count and amount reconcile to the selected summary value.

A monthly margin chart has 24 points, and finance users want exact revenue, cost, margin, and margin percent on hover without four permanent labels. How would you implement it?

I would use a concise tooltip so the chart stays readable while exact values remain available on demand. - Add Month, Revenue, Cost, Margin, and Margin Percent to the tooltip content with consistent currency and percentage formats. - Order the 5 fields by business importance and label each one explicitly instead of exposing technical column names. - Hover 3 known months, reconcile all tooltip values to a validation table, and check that the content remains legible at the published report size.

A Power BI page has a revenue bar chart and a product table, but clicking one bar should filter the table while leaving a company-total KPI card unchanged. How would you author that behavior?

I would configure visual interactions rather than changing the underlying measures. - Select the bar chart, open Edit interactions, set the product table to Filter, and set the company-total card to None. - Keep the interaction choices visible in the report design checklist because the unchanged card is an intentional requirement, not a broken filter. - Click 2 bars and verify that the table row count changes while the KPI card remains equal to the company total.

A Looker dashboard has 6 tiles from 2 Explores, and 15 managers need one Region filter to affect every compatible tile. How would you set it up?

I would add one dashboard Region filter and map it to the correct region field for each tile. - Choose a user-friendly control and leave it unfiltered by default when all regions should appear, rather than hard-coding a region inside each query. - For every one of the 6 tiles, map the dashboard filter to the field from that tile's Explore when the field is available. - Select 2 regions and verify each tile's query result, documenting any tile that cannot be mapped instead of implying it is filtered.

Sixty field managers view a Power BI operations page on phones, but the desktop page has 10 visuals and a wide matrix. What mobile layout would you create?

I would create a dedicated phone layout that prioritizes the few visuals needed in the field. - Place 3 KPI cards first, followed by 1 alert chart and a narrow action table, instead of shrinking all 10 desktop visuals. - Leave the wide matrix out of the phone layout and provide essential columns through a mobile-friendly table or a drill-through page, and size controls for touch rather than precise mouse input. - Preview the phone layout at 2 common device sizes and confirm that the top decision can be made without horizontal scrolling.

A dashboard for 120 employees uses red and green status dots, small labels, and keyboard users must reach 7 visuals. What accessibility changes would you make before handoff?

I would make status understandable without color and ensure every visual has readable, navigable context. - Add text or shape cues such as On target and Below target, and check at least a 4.5 to 1 contrast ratio for normal-sized text. - Set meaningful titles and alt text for all 7 visuals, use readable font sizes, and avoid labels that depend only on hover. - Review tab order with a keyboard and test the page at 200 percent zoom plus a grayscale view before handoff.

BI Developer interview questions

100 real questions with model answers and explanations for Junior candidates.

See a BI Developer resume example →

Practice with flashcards

Spaced repetition · Hunter Pass

Questions

dashboardsci-cd

I would first identify the decision the manager must make from the dashboard.

Ask whether the Friday meeting reallocates leads, reviews quota risk, or follows overdue deals, because each decision needs a different primary KPI.
Define the audience and cadence as 1 manager plus 18 representatives using a daily snapshot, then record which owner, stage, and region fields they need to filter.
Leave the meeting with a 1-page wireframe and 3 signed examples of deals that should appear as needing attention, then use those rows as acceptance checks.

Why interviewers ask this: The interviewer is evaluating whether you turn a vague dashboard request into a specific decision, audience, and testable artifact.

dashboardsresiliencekpi

I would write a metric contract for monthly net revenue before creating the KPI card.

Set the formula to paid invoice amount minus posted refunds, exclude canceled invoices, and state whether the month comes from payment date or invoice date.
Fix the display grain at 1 calendar month and document currency conversion, tax treatment, and the cutoff time for the daily refresh.
Reconcile 3 sample months to the finance ledger and require a zero unexplained difference before the card is accepted.

Why interviewers ask this: The interviewer checks whether you define status, time, and reconciliation rules instead of placing an ambiguous sum on a dashboard.

funnelkpidashboards

I would define the KPI at one user per funnel step per calendar day, not one raw event row.

Deduplicate the 800,000 events by user_id, event_date, and step so 5 checkout attempts by 1 user count once for that step.
Calculate daily conversion as distinct purchasers divided by distinct eligible funnel entrants, with both populations using the same cohort rule.
Validate the grain with a 20-row user journey sample and add a grain note beside the dataset definition.

Why interviewers ask this: The interviewer is testing whether you can prevent repeated events from inflating a dashboard KPI by defining grain explicitly.

dashboardschartscss

I would build a summary page that shows status first and supporting trends second.

Put 3 KPI cards for revenue, margin, and target attainment in the first row, each with the current value and variance to target.
Use 1 monthly line chart for the 12-month trend and 1 sorted bar chart for the 6 regions, because executives need direction and comparison rather than row-level detail.
Keep the page to about 5 visuals, place the latest refresh timestamp visibly, and test that all 3 questions can be answered within a 5-minute walkthrough.

Why interviewers ask this: The interviewer evaluates whether chart choice and information density match a time-constrained executive audience.

I would use a line chart with a separate target reference line.

Put week on a continuous x-axis and qualified lead count on the y-axis so the 26-week direction and seasonality remain visible.
Add a clearly labeled horizontal line at 400 rather than a second bar series, which makes misses and recoveries easy to scan.
Check the chart at 1366 pixels wide and verify that all 26 points are readable without labels on every mark.

Why interviewers ask this: The interviewer checks whether you select a visual that communicates trend and target variance without clutter.

chartsdashboards

I would replace the pie with a descending horizontal bar chart.

Sort all 14 categories by ticket count so the largest workload is visible without comparing angles or hunting through a legend.
Show direct count labels and optionally group categories below 1 percent into Other only if the category detail remains available elsewhere.
Test the visual with the longest category name and a 1280-pixel screen to ensure labels do not truncate or overlap.

Why interviewers ask this: The interviewer evaluates whether you can replace a poor chart with a comparison visual that handles many categories and long labels.

kpi

I would use one KPI card with the actual, target variance, and prior-day change shown together.

Display 92 percent as the primary value and a labeled minus 3 percentage-point variance to the 95 percent target.
Add the prior-day comparison as a secondary value, such as down 1.2 percentage points, without mixing percentage points with percent change.
Use text or an icon in addition to color and verify the card remains understandable in grayscale and at 200 percent zoom.

Why interviewers ask this: The interviewer checks whether you distinguish target variance from period change and communicate status accessibly.

dashboardscharts

I would lead with an actionable table because users must identify specific customer rows.

Include customer, renewal date, owner, contract value, days remaining, and last-contact date, then sort by days remaining ascending.
Add filters for the 30 owners and renewal-value band, while keeping a small card above the table for total value due within 14 days.
Validate the table with 10 known accounts and check that the first screen shows enough columns without horizontal scrolling at the standard laptop width.

Why interviewers ask this: The interviewer evaluates whether the chosen visual supports row-level operational action rather than decorative aggregation.

dashboardskpi

I would not merge the two definitions into one unlabeled KPI.

Document 2 candidate definitions with formulas, date fields, and example customer IDs, then ask the dashboard owner to approve the business meaning.
If both decisions matter, show 2 explicitly named measures such as Purchasing active, 90 days and Product active, 30 days rather than hiding the difference.
Test 10 boundary customers against both rules and attach the approved metric contract before the 4-day build starts.

Why interviewers ask this: The interviewer checks whether you surface conflicting business definitions and preserve semantic clarity under deadline pressure.

dashboardswarehouse

I would scope the first version to the supervisors' daily stock-risk decision and make its acceptance criteria explicit.

Limit version 1 to 3 KPIs, current stock, items below reorder point, and stock value, plus 1 warehouse filter and 1 item-level table.
Write a wireframe and a field mapping from each visual to the source column or measure before authoring the dashboard.
Agree on 5 acceptance checks, including known low-stock SKUs from at least 2 warehouses, and complete a review with 2 supervisors before the 3-day deadline.

Why interviewers ask this: The interviewer evaluates whether you can control scope while still delivering a testable BI artifact for a defined audience.

bipower-bicharts

I would use visual-level, page-level, and report-level filters for those three scopes.

Put Region equals West on the single chart's visual filter so the other 11 visuals keep their own region context.
Put Year equals 2026 on the relevant page filter so every visual on that page receives it.
Put Status equals Active on the report-level filter, then inspect 1 visual on each of the 4 pages to confirm the filter scope.

Why interviewers ask this: The interviewer checks whether you understand Power BI filter scopes and can verify their effect without overfiltering the report.

configbipower-bi

I would add clearly labeled slicers bound to dimension fields and keep their state easy to inspect.

Use a single-select Year slicer when the KPI specification allows only 1 year, and multi-select slicers for the 5 regions and 3 product families.
Place the slicers consistently at the top of the report canvas and show the current selection in a subtitle or filter summary.
Test 3 combinations against a simple table of expected totals and use Edit interactions if a visual must intentionally ignore one slicer.

Why interviewers ask this: The interviewer evaluates basic Power BI slicer authoring, interaction scope, and validation against expected totals.

chartsdashboardsbi

I would create a Tableau filter action from the region chart to the detail table.

Set the source sheet to the 8-region bar chart, run the action on Select, and target only the order-detail sheet.
Map Region to Region and choose the clearing behavior deliberately, such as Show all values when the selection is cleared.
Click 3 regions and compare the target row counts with worksheet filters to confirm that the action passes the intended field.

Why interviewers ask this: The interviewer checks whether you can connect Tableau worksheets with a correctly scoped dashboard filter action.

bilooker

I would build the query from governed fields in the existing Explore, verify it, and save it as a dashboard tile.

Select Week and Product Category as dimensions plus Orders and Revenue as measures, then filter to the required date range.
Run the query and check 3 category totals against known source results before changing visualization settings.
Save the result to the shared dashboard with a clear title and preserve the field definitions from LookML rather than recreating KPI logic in the tile.

Why interviewers ask this: The interviewer evaluates whether you can use a Looker Explore and its modeled measures to author a reliable dashboard tile.

configbipower-bi

I would create a dedicated order-detail page and add Region to its drill-through filter well.

Put an order table on the target page with order ID, date, customer, and amount, and add a Back button for the return path.
Users can right-click a regional data point and choose the drill-through page, which carries that region's filter context.
Test 2 regions and confirm both the detail row count and amount reconcile to the selected summary value.

Why interviewers ask this: The interviewer checks whether you know the required Power BI drill-through field, navigation path, and reconciliation check.

chartscss

I would use a concise tooltip so the chart stays readable while exact values remain available on demand.

Add Month, Revenue, Cost, Margin, and Margin Percent to the tooltip content with consistent currency and percentage formats.
Order the 5 fields by business importance and label each one explicitly instead of exposing technical column names.
Hover 3 known months, reconcile all tooltip values to a validation table, and check that the content remains legible at the published report size.

Why interviewers ask this: The interviewer evaluates whether tooltips add useful detail without turning the main visual into a field dump.

chartsbipower-bi

I would configure visual interactions rather than changing the underlying measures.

Select the bar chart, open Edit interactions, set the product table to Filter, and set the company-total card to None.
Keep the interaction choices visible in the report design checklist because the unchanged card is an intentional requirement, not a broken filter.
Click 2 bars and verify that the table row count changes while the KPI card remains equal to the company total.

Why interviewers ask this: The interviewer checks whether you can control Power BI cross-filtering at the visual level and verify intentional exceptions.

bilookerdashboards

I would add one dashboard Region filter and map it to the correct region field for each tile.

Choose a user-friendly control and leave it unfiltered by default when all regions should appear, rather than hard-coding a region inside each query.
For every one of the 6 tiles, map the dashboard filter to the field from that tile's Explore when the field is available.
Select 2 regions and verify each tile's query result, documenting any tile that cannot be mapped instead of implying it is filtered.

Why interviewers ask this: The interviewer evaluates whether you understand that Looker dashboard filters must be connected to compatible fields per tile.

bipower-bimatrix

I would create a dedicated phone layout that prioritizes the few visuals needed in the field.

Place 3 KPI cards first, followed by 1 alert chart and a narrow action table, instead of shrinking all 10 desktop visuals.
Leave the wide matrix out of the phone layout and provide essential columns through a mobile-friendly table or a drill-through page, and size controls for touch rather than precise mouse input.
Preview the phone layout at 2 common device sizes and confirm that the top decision can be made without horizontal scrolling.

Why interviewers ask this: The interviewer checks whether you treat mobile as a prioritized layout rather than a compressed desktop canvas.

dashboardsa11y

I would make status understandable without color and ensure every visual has readable, navigable context.

Add text or shape cues such as On target and Below target, and check at least a 4.5 to 1 contrast ratio for normal-sized text.
Set meaningful titles and alt text for all 7 visuals, use readable font sizes, and avoid labels that depend only on hover.
Review tab order with a keyboard and test the page at 200 percent zoom plus a grayscale view before handoff.

Why interviewers ask this: The interviewer evaluates practical dashboard accessibility through color, text alternatives, keyboard order, and zoom checks.

Locked questions

21
You receive 1.2 million sales rows plus customer, product, and store attributes for a revenue dashboard. Which tables would be facts and which would be dimensions?
dashboardsmodeling
22
An orders source has 250,000 order lines, but the requested dashboard shows order count, units, and revenue by product. What grain should FactSales use?
dashboardsgrain
23
A Power BI model has 5,000 customers and 900,000 sales rows. How would you configure the customer-to-sales relationship for a regional KPI dashboard?
configbipower-bi
24
A daily revenue dashboard covers 2024 through 2026 and needs month, quarter, weekday, and fiscal-year filters even on dates with no sales. What should the date dimension contain?
modelingdashboards
25
An order dashboard needs to analyze 400,000 rows by both order date and ship date. How would you model the two date roles without creating ambiguous filtering?
dashboards
26
During today's load, 230 sales rows reference product codes not yet present in the product dimension, but the revenue dashboard must still preserve their value. How would you handle them?
dashboardsmodeling
27
Two source systems both use product code 1001 for different products, and 3 million sales rows must map to one DimProduct. Would you use product code alone as the dimension key?
modelingsystem-design
28
A shipping dashboard needs order number for search and detail, but each order number can repeat across several line rows and has no descriptive attributes. Would you create a separate dimension?
dashboardsmodeling
29
A warehouse dashboard receives one end-of-day inventory balance for each of 4,000 products across 30 days. What fact grain and aggregation warning would you document?
aggregationwarehousegrain
30
A monthly dashboard compares daily actual sales with one budget row per product category per month. Should you join both datasets into one fact table?
joinsmodelingdashboards
31
A BI query must list all 10,000 customers and their 2026 revenue, including customers with zero orders. Which SQL join would you use?
sqljoinsqueries
32
Joining 600,000 order lines directly to 40,000 payment rows doubles revenue on a dashboard because some orders have multiple payments. How would a CTE help?
ctesqldashboards
33
A dashboard needs one row per month and region from 2 million sales rows, with revenue and order count. What SQL aggregation would you write?
sqlaggregationdashboards
34
A customer-status table contains 700,000 updates, and a dashboard needs only the latest status for each of 85,000 customers. How would you use a window function?
window-functionssqldashboards
35
A finance table has 36 monthly revenue rows across 2024 through 2026, and the dashboard needs a running total that restarts each January. How would you use a SQL window function?
sqlwindow-functionsdashboards
36
Power Query imports 150,000 CSV rows where order dates are text, revenue uses decimal commas, and 42 customer IDs have surrounding spaces. What cleanup steps would you apply?
queriespower-query
37
A Power Query model reads 5 million SQL rows but needs only 2026 data and 8 columns for a dashboard. How would you preserve query folding?
sqlqueriesschema
38
Power Query has 2,500 product rows and an 18-row category lookup, and DimProduct needs category name alongside each product's brand. How would you merge them?
queriespower-query
39
You must create a dbt staging model for 900,000 raw order rows before a Power BI deadline tomorrow. What belongs in that staging model?
bidbtpower-bi
40
A dbt sales mart joins staged orders to customers for a 20-user dashboard. How would you reference models and check the join before handoff?
joinsdbtdashboards
41
A Power BI report has 2 million sales rows and needs margin percent to change with year, region, and product slicers. Should you use a measure or calculated column?
measuresdaxbi
42
Each of 500,000 sales rows has Quantity and UnitPrice, and the model needs a stored LineAmount for grouping into fixed price bands after refresh. How would row context apply?
dax
43
The Revenue measure in a Power BI report shows 8.4 million for all data, 2.1 million after selecting West, and 600,000 after also selecting March. Why does the same measure return different values?
bipower-bimeasures
44
A margin KPI must divide 320,000 margin by 2,000,000 revenue, but some product categories have zero revenue. Which DAX pattern would you use?
kpidaxcss
45
A customer dashboard contains 1.8 million order rows from 72,000 customers, and it needs unique buyers by month. Which DAX aggregation would you use?
dashboardsaggregationdax
46
A 12-user service dashboard needs a Closed Tickets KPI that responds to date and team slicers but always includes only status Closed. How would you write it with CALCULATE?
dashboardsfilterskpi
47
A Power BI page for 30 managers needs a title that says Sales: All regions with no Region filter, Sales: West with exactly one region, and Sales: Multiple regions with two or more. How would you build the title measure?
bipower-bimeasures
48
A monthly revenue dashboard needs prior-year revenue for January through June 2026, and the model has a continuous marked date table. How would you calculate the comparison?
dashboardsdax
49
A KPI page shows 24 monthly order totals and needs the change versus the immediately previous month under a product slicer. How would you build it in DAX?
kpidaxfilters
50
A finance dashboard needs year-over-year growth from 6.0 million to 7.5 million and must show blank when prior-year revenue is missing. How would you compose the DAX measures?
measuresdaxdashboards
51
A Power BI matrix shows 1,240 orders across its rows, but the total says 980. How would you investigate before changing the measure?
bipower-bimatrix
52
A Power BI sales table has 3 lines with quantities and unit prices of 2 at $10, 1 at $50, and 3 at $20. A measure multiplies SUM of quantity by SUM of price and returns $480. How would you correct it?
bipower-bimeasures
53
A regional conversion table shows 42%, 38%, and 52%, while the total displays 132%. How would you fix the report?
54
A year-to-date sales card stays at $3.2 million when the user changes the date slicer's upper bound from June 30 to May 31. How do you debug it?
filters
55
A column chart makes the $102,000 column look three times as tall as the $98,000 column because its axis starts at $96,000. What would you change?
chartsschema
56
A dual-axis chart overlays 1.8 million visits and a 2.4% conversion rate, making both lines appear perfectly correlated. How would you redesign it?
chartsconversion
57
Users select Q2 in a slicer, but the report title still says Q1 Results and an axis says Revenue in thousands although values are in millions. What do you do?
filters
58
Your executive dashboard reports $1.24 million monthly revenue, while Finance reports $1.19 million. How would you reconcile the KPI?
dashboardskpireconciliation
59
You inherit a Power BI page with 12 visuals and 4 slicers. What dashboard QA would you perform before sharing it with 35 sales managers?
dashboardsfiltersbi
60
A dashboard passes its normal view, but selecting a new region produces 6 blank visuals and one misleading zero KPI. How would you handle the edge case?
dashboardskpi
61
A Power BI Service refresh that normally takes 18 minutes now fails after 45 minutes. What are your first steps?
bipower-bi
62
Power Query reports The column CustomerName was not found after a 250,000-row source file changed from 12 to 13 columns. How do you repair it safely?
queriesschemapower-query
63
A Power Query refresh produces 3,400 conversion errors after a CSV switches dates from MM/DD/YYYY to DD/MM/YYYY. What would you do?
queriespower-query
64
A Power BI report is built against SQL-DEV, but its published semantic model must refresh from SQL-PROD without editing M code for every release. How would you configure the switch?
configbipower-bi
65
You must build a report over 8 million sales rows that updates four times daily. How would you choose between Import and DirectQuery?
66
A DirectQuery report over 15 million rows takes 14 seconds whenever a user changes a slicer. What would you check before switching to Import?
filters
67
A scheduled Power BI refresh fails with a gateway offline message at 06:00, while Desktop refresh still works. How do you troubleshoot it?
gatewaybipower-bi
68
A Power BI Service refresh begins failing with invalid credentials one day after a database password rotation. What actions do you take?
bipower-bidatabase
69
A report reads a 60,000-row Excel file from a developer's laptop and cannot refresh after publication. How would you make refresh reliable?
excel
70
Power BI refresh history says Completed in 11 minutes, but the report still shows data only through July 13 instead of July 15. What do you check?
bipower-bi
71
A Power BI page with 24 visuals opens in 19 seconds. How would you use Performance Analyzer to improve it?
profilingbipower-bi
72
Performance Analyzer shows that a Power BI table loading 250,000 transaction rows is the main page delay, but managers only need the 50 largest overdue orders. How would you redesign it?
profilingbipower-bi
73
A 6-million-row Power BI model is 1.7 GB, and the fact table includes unique Transaction ID and 200-character Comment columns. What would you change?
transactionsschemamodeling
74
A report became slower and started showing ambiguous totals after you changed 4 relationships to bidirectional filtering. How would you recover?
75
You created East and West RLS roles for 120 sales users. How do you test that an East user cannot see West revenue?
row-level-security
76
A new sales user sees a blank report after being assigned to the North RLS role. What do you investigate?
row-level-security
77
You need to release a Power BI report to 65 viewers while only 4 developers should edit workspace content. How would you share it?
bipower-bi
78
A report shows only regional totals, but users can export 90,000 underlying customer rows including email addresses. What would you do?
79
A teammate suggests Publish to web so 300 contractors can open a Power BI report without licenses. How do you respond?
bipower-bi
80
A viewer requests Build permission on a semantic model used by 40 people so they can create an Excel pivot. What checks do you make?
semantic-layerexcel
81
You fixed 3 measures in a Power BI PBIX, but production users need the old version available if totals regress. What publish workflow do you use?
bipower-bimeasures
82
You must change a LookML field used by 18 dashboards. What Git workflow should a junior BI developer follow?
gitbidashboards
83
Eight Power BI authors each copied the same 22-measure model into a separate PBIX, and revenue fixes now require 8 manual updates. How would you introduce governed self-service?
bipower-bimeasures
84
What deployment checklist would you use for a 9-page Power BI report scheduled to refresh at 07:00?
deploymentbipower-bi
85
Twelve sales managers must perform UAT on a new pipeline dashboard before Friday. How would you organize the test?
dashboardsuatci-cd
86
A Power BI report opens with Region set to West and Year set to 2025 because the author saved that test state, while the approved default is all regions and 2026. How would you fix and verify the default before release?
bipower-bi
87
A Power BI phone layout is already built for 70 field staff, but a tester cannot clear the Region slicer and a drill-through page has no obvious return path. What mobile QA and fixes would you perform?
filtersdrill-throughbi
88
A Power BI app was sent to 80 employees, but usage metrics show only 14 unique viewers in 30 days. What do you check first?
monitoringbipower-bi
89
The primary BI developer leaves on Friday, and a 9-page Power BI report used by 45 people depends on 4 sources, 1 gateway, and a daily refresh. What handoff would you require?
gatewaybipower-bi
90
After a report is renamed and moved into a new Power BI app section, usage drops from 55 to 18 weekly users. What would you do?
bipower-bi
91
A Tableau workbook uses an extract refreshed nightly, but users expect today's 10:00 orders to appear immediately. How do you explain and resolve the issue?
bitableau
92
A Tableau workbook switched from a 20-minute extract refresh to a live connection, and each filter now takes 12 seconds. What do you investigate?
bitableau
93
A Tableau Cloud extract has 900,000 rows, but 120 corrected orders do not change after an incremental refresh keyed on Order ID. Why, and what would you do?
bitableaurefresh
94
A Looker Explore shows 6,200 customers, but the approved customer dashboard shows 5,850. How would you validate the Explore?
bilookervalidation
95
A Looker dashboard still shows yesterday's $1.2 million revenue after the warehouse has today's approved rows. How would you determine whether query caching is the cause?
querieswarehousebi
96
You rename a LookML field and Content Validator reports 23 broken Looks and dashboard tiles. What do you do?
dashboardsvalidationbi
97
You find 4 nearly identical pipeline reports with 22, 17, 6, and 0 weekly users. How would you reduce duplication?
ci-cd
98
A support ticket says The dashboard is wrong and includes only a screenshot showing 0 orders. How do you triage it?
dashboards
99
A Power BI app update at 09:00 makes 5 KPI cards blank for 140 users. What is your incident response?
bipower-biincidents
100
11 Looker dashboard tiles fail minutes before an executive meeting, and one tile's generated SQL says the Looker warehouse account lacks SELECT on analytics.orders. How would you recover?
sqlwarehousebi

Questions

1A sales manager asks for a pipeline dashboard for 18 representatives by Friday but only says it should show what needs attention. What would you clarify first?

2A finance dashboard has 12,000 invoice rows, including cancellations and partial refunds, and the CFO asks for one monthly revenue KPI. How would you define it before building the card?

3A product funnel dashboard uses 800,000 event rows, and one user can trigger checkout_started five times. What grain would you choose for a daily conversion KPI?

4You have one page for 6 executives who review revenue, margin, and target attainment in 5 minutes each Monday. How would you choose the dashboard layout and charts?

5A marketing lead wants to compare weekly qualified leads for the last 26 weeks against a target of 400. Which visual would you use and why?

6A support dashboard must compare 14 issue categories by ticket count for 25 team leads. The current pie chart has unreadable slices. What would you replace it with?

7An operations director needs to see whether yesterday's 92 percent on-time delivery KPI met a 95 percent target and how it changed from the prior day. How would you present it?

8Thirty account managers need a daily dashboard to find customers with renewals due in the next 14 days, including owner, value, and last contact. Would you lead with a chart or a table?

9Sales defines an active customer as anyone with an order in 90 days, while Customer Success uses a login in 30 days. Both teams want one active customers KPI on a shared dashboard in 4 days. What would you do?

10You have 3 days to deliver a first inventory dashboard from 2 million stock-movement rows for 8 warehouse supervisors. How would you define a junior-sized first version?

11A Power BI report has 4 pages and 12 visuals, but the regional manager wants West applied only to one chart, 2026 applied to one page, and Active applied everywhere. Where would you place each filter?

12A Power BI page serves 40 sales users who need to switch among 5 regions and 3 product families while keeping the selected year visible. How would you configure the slicers?

13A Tableau dashboard has a bar chart for 8 regions and a detail table with 60,000 order rows. Users want a click on a region to narrow the table. How would you build it?

14In Looker, 12 product managers need a weekly table of orders and revenue by category from an existing Explore. What authoring steps would you take?

15A Power BI summary page shows sales for 20 regions, and users must open a page listing the orders behind the selected region. How would you configure drill-through?

16A monthly margin chart has 24 points, and finance users want exact revenue, cost, margin, and margin percent on hover without four permanent labels. How would you implement it?

17A Power BI page has a revenue bar chart and a product table, but clicking one bar should filter the table while leaving a company-total KPI card unchanged. How would you author that behavior?

18A Looker dashboard has 6 tiles from 2 Explores, and 15 managers need one Region filter to affect every compatible tile. How would you set it up?

19Sixty field managers view a Power BI operations page on phones, but the desktop page has 10 visuals and a wide matrix. What mobile layout would you create?

20A dashboard for 120 employees uses red and green status dots, small labels, and keyboard users must reach 7 visuals. What accessibility changes would you make before handoff?

Locked questions