Difficulty: Hard
Given two dashboard averages with different group sizes, how should their overall mean be calculated?
A Use the larger mean
B Weight by group size
C Choose the newest mean
D Average the two means equally
Answer: B. Weight by group size. A weighted mean accounts for each group’s contribution to the total.
Why must a fact table have a declared grain?
A Reports need matching colors
B Measures use consistent events
C Refreshes need no filters
D Dimensions require equal row counts
Answer: B. Measures use consistent events. A declared grain defines the event represented by each row and protects calculations.
What can a many-to-many bridge table help represent?
A Associations between multiple entities
B Refresh timing by date
C A dashboard’s color palette
D A single numeric measure
Answer: A. Associations between multiple entities. A bridge table models associations when entities can relate to many others.
How does Type 2 slowly changing dimension preserve history?
A Insert a new versioned row
B Move data to staging only
C Delete related facts
D Overwrite the old attribute
Answer: A. Insert a new versioned row. Type 2 adds a new dimension version so historical facts retain prior context.
Why recompute a ratio after aggregating its groups?
A Filters cannot change ratios
B Combined totals may differ
C Ratios are always additive
D Averages equal totals
Answer: B. Combined totals may differ. Ratios should usually be recalculated from aggregated components.
What should a BI security test verify for each role?
A All users see every table
B Only permitted rows are visible
C Refresh logs contain no dates
D Dimensions have equal sizes
Answer: B. Only permitted rows are visible. Role testing confirms row-level rules work as intended.
When a late-arriving fact lacks its dimension record, what helps preserve reporting?
A Change its measure to zero
B Rename the source table
C Delete the transaction
D Use a temporary unknown key
Answer: D. Use a temporary unknown key. A controlled unknown member lets the fact load until dimension data arrives.
What does schema drift mean in an ETL pipeline?
A A KPI exceeds its target
B A dimension is sorted alphabetically
C Source structure changes unexpectedly
D A report page is hidden
Answer: C. Source structure changes unexpectedly. Schema drift occurs when source fields or types change and downstream mappings may break.
How should an incremental pipeline avoid duplicated loads after retries?
A Make writes idempotent
B Disable source validation
C Append every batch blindly
D Delete all history first
Answer: A. Make writes idempotent. Idempotent processing lets retries avoid applying the same data twice.
Why can a full reload cause risk for large historical models?
A It reduces table count
B Adds processing time and resources
C It removes all source errors
D It guarantees fresher data immediately
Answer: B. Adds processing time and resources. Full refreshes process the full dataset and can consume more time and capacity.
What does a KPI denominator definition control?
A Dimension row order
B Dashboard color selection
C Data source ownership
D Meaning of the reported rate
Answer: D. Meaning of the reported rate. Changing the denominator changes the business interpretation of a rate.
How can sampling bias affect an executive dashboard?
A Improve every forecast automatically
B Guarantee accurate causation
C Underrepresent important populations
D Remove all outliers
Answer: C. Underrepresent important populations. Biased samples can make reported patterns unrepresentative of the broader population.
Why can dual-axis charts mislead viewers?
A Scales can exaggerate visual relationships
B They remove data lineage
C They guarantee equal units
D They prevent time filtering
Answer: A. Scales can exaggerate visual relationships. Independent axis scales can make unrelated movements appear strongly aligned.
What should a data validation rule test before loading facts?
A Dashboard subscription status
B User profile image
C Required keys and measure ranges
D Chart font size
Answer: C. Required keys and measure ranges. Validation checks fields, keys, and value constraints before data enters analysis.
Why can changing a slicer alter a KPI’s denominator?
A The dimension table vanishes
B The source system reboots
C The report title is updated
D Filters change included rows
Answer: D. Filters change included rows. Filters can change both numerator and denominator populations.
What does a late-arriving dimension record require?
A Ignore the business key
B Rebuild all dashboards manually
C Delete the event permanently
D Update the unknown member relationship
Answer: D. Update the unknown member relationship. Late dimension data may require matching facts to a resolved dimension record.
Which design choice improves consistent metric reuse?
A Store definitions in email
B Centralize measures in semantic model
C Copy formulas into every report
D Use unlabeled spreadsheet columns
Answer: B. Centralize measures in semantic model. Centralized measures reduce conflicting calculation definitions across reports.
How can time-zone mismatches distort daily reporting?
A Measures become non-additive automatically
B Dimensions lose their keys
C Events fall into different dates
D Refreshes stop all visual filters
Answer: C. Events fall into different dates. Timezone conversion can move events across date boundaries.
Why should dashboards avoid excessive visuals?
A Too much information weakens focus
B One page per KPI
C Filters become mathematically incorrect
D More charts guarantee better decisions
Answer: A. Too much information weakens focus. A dashboard should emphasize the most useful signals instead of overwhelming users.
What should lineage record across data preparation?
A Only final chart colors
B Report page order
C Source, transformations, and destination
D User login frequency
Answer: C. Source, transformations, and destination. Lineage tracks where data came from and how it was transformed.
When import mode is used, where is model data analyzed?
A Inside each report image
B Only in the source application
C In the imported semantic model
D In a dashboard tile file
Answer: C. In the imported semantic model. Import mode loads a copy of data into the semantic model for analysis.
What is a key DirectQuery tradeoff?
A Source queries; performance may vary
B Fixed answers without refresh
C Automatic historical archiving
D No need for data relationships
Answer: A. Source queries; performance may vary. DirectQuery queries the source at interaction time, making performance depend on source and network.
Why monitor refresh failures as a governance issue?
A They improve source quality automatically
B They change chart types
C They create new dimensions
D Users may trust stale measures
Answer: D. Users may trust stale measures. Failed refreshes can leave published analyses outdated and mislead decisions.
Which governance control clarifies who owns a shared metric?
A Unlabeled chart legend
B Hidden source query
C Random refresh time
D Named business owner
Answer: D. Named business owner. A named owner is accountable for definition, review, and change control.
What makes a KPI comparison fair across departments?
A Consistent scope and calculation rules
B Equal staff headcount
C Identical source applications
D Same chart color
Answer: A. Consistent scope and calculation rules. Comparable KPIs need the same population, time period, and calculation definition.