Enhancement Request: Give Users Greater Control Over Bing Search Suggestions Using Context-Aware Personalization and Privacy Controls

Analytics types

• Descriptive: What happened? Summarizes past results.
• Diagnostic: Why did it happen? Finds causes or explanations.
• Predictive: What might happen? Forecasts likely outcomes.
• Prescriptive: What should we do? Recommends an action.

Power BI components

• Power BI Desktop: Connect/transform data, build models, write DAX, create reports.
• Power BI Service: Online publishing, sharing, collaboration, apps, refresh and workspace management.
• Power BI Mobile: View/interact with content on mobile devices.
• Power Query: Connect, clean, reshape and transform data before modeling.
• DAX: Formula language for model calculations.
• Semantic model: Data, relationships, measures and model metadata used by reports.
• Report: One or more interactive pages; usually built on a semantic model.
• Dashboard: One-page canvas in Power BI Service, assembled from pinned tiles.
• Workspace: Collaborative area for managing Power BI content.
• App: Curated content package distributed to consumers.
• Microsoft Fabric: Integrated analytics platform across data engineering, data science, real-time analytics, warehousing and BI.
• Copilot: AI assistance for supported authoring/analysis tasks; availability depends on capacity, tenant settings and permissions.

MODULE 2 — PREPARE DATA

Power Query and data quality

• Power Query Editor: Interface for data cleaning and transformation.
• Applied Steps: Recorded transformation steps; steps run in sequence.
• Data profiling: Helps inspect quality and distribution.
  • Column quality: Valid, error and empty values.
  • Column distribution: Distinct/unique values and distribution.
  • Column profile: Detailed statistics/profile for a column.
• Common problems: nulls, errors, duplicate rows, inconsistent spelling/case, wrong data types, mixed formats, irrelevant rows/columns and outliers.
• Set correct data types; use consistent categories; remove or handle errors/nulls based on requirements; verify results after transformations.

Common transformations

• Filter rows: Keep rows meeting a condition.
• Remove rows: Delete unwanted records.
• Choose/remove columns: Reduce unnecessary fields.
• Change type: Convert text, date, whole number, decimal, etc.
• Split column: Divide one column using delimiter/position.
• Replace values: Standardize inconsistent values.
• Group By: Aggregate rows into groups.
• Pivot: Turn distinct values from rows into columns.
• Unpivot: Turn columns into attribute-value rows.
• Append queries: Stack rows from tables with compatible columns.
• Merge queries: Join tables using matching key columns.
• Transpose: Swap rows and columns.

Merge vs Append

• Merge = join side-by-side using matching key(s); similar to a database join.
• Append = stack vertically by adding rows from another table. MCQ: Combine customer details using CustomerID → Merge. Combine January and February sales rows → Append.

Query folding vs data reduction

• Query folding: Power Query translates supported steps into a query executed by the source system. Can improve efficiency by pushing work to the source.
• Data reduction: Use fewer rows/columns or summarize data to reduce volume and model size.
• Query folding is not the same as reducing data. Folding concerns where transformations execute; reduction concerns how much data is kept.
• Some transformations or sources may prevent folding; check folding indicators/source query when available.

Storage/connectivity modes

• Import: Data is loaded into the model; typically fast report interaction, but data needs refresh.
• DirectQuery: Queries the source when users interact; source performance and connectivity matter.
• Direct Lake: In supported Fabric scenarios, works with data in OneLake without the usual import-copy approach.
• Live connection: Connects to an existing semantic model/Analysis Services model with limited local modeling compared with importing data.
• Composite model: Combines supported storage modes/sources in one model. MCQ: Fast local model + scheduled refresh → Import. Query source at interaction time → DirectQuery.

M vs DAX

• M: Power Query language; used for data preparation before loading.
• DAX: Used for calculations in the semantic model and reports. Remember: Clean/reshape → Power Query/M. Measures and model calculations → DAX.

---

MODULE 3 — MODEL DATA & DAX

Measures, columns and tables

• Measure: Calculated at query time and responds to filter context; ideal for totals, ratios and KPIs.
• Calculated column: Computes a value for each row and stores it in the model; can increase model size.
• Calculated table: Creates a table through a DAX expression during model processing/refresh.
• Visual calculation: Calculation defined on a visual's displayed data; useful for visual-level calculations such as running totals. MCQ clues: Responds to slicers and report filters → measure. Stored value per row → calculated column. Creates a new model table → calculated table.

Core DAX functions

• SUM(Column): Adds values in one numeric column.
• AVERAGE(Column): Mean of numeric values.
• COUNT(Column): Counts nonblank values in a column.
• COUNTROWS(Table): Counts rows in a table.
• DISTINCTCOUNT(Column): Counts distinct values.
• MIN/MAX: Smallest/largest value.
• DIVIDE(a,b): Division with optional alternate result for divide-by-zero.
• SUMX(Table, expression): Iterates rows, evaluates expression per row, then sums results.
• RELATED(Column): Retrieves a related value from the one-side/related table when relationship context supports it.
• LOOKUPVALUE: Searches for a value based on matching search conditions; relationships are often preferable when suitable.
• VALUES(Column): Returns distinct values visible in the current context (including behavior for blank/relationship cases).
• SELECTEDVALUE(Column, alternate): Returns the value when exactly one distinct value is selected; otherwise returns alternate/blank.
• FILTER(Table, condition): Returns rows that meet a condition.
• CALCULATE(expression, filters...): Evaluates an expression under modified filter context.
• CALCULATETABLE(table expression, filters...): Returns a table under modified filter context.
• REMOVEFILTERS(Table/Column): Removes specified filters when used in a filter-modifying expression.
• ALL(Table/Column): Returns all rows/values ignoring filters on the specified table/column in relevant contexts; can also act as a filter modifier.
• KEEPFILTERS(condition): Intersects a new filter with existing filters rather than simply replacing filters on the same columns.

SUM vs SUMX

• SUM(Sales\[Amount]) adds an existing numeric column.
• SUMX(Sales, Sales\[Quantity] \* Sales\[UnitPrice]) calculates an expression per row, then sums it. MCQ clue: “Multiply quantity by price for every row, then total” → SUMX.

MODULE 4 — TIME INTELLIGENCE, VISUAL CALCULATIONS & PERFORMANCE

Date table and time intelligence

• A proper Date/Calendar table contains one row per date, continuous dates, and appropriate date attributes.
• Marking a date table helps Power BI identify the date column for time-intelligence scenarios.
• Time-intelligence calculations commonly compare periods such as year-to-date, previous year, or same period last year.
• A correct date table and relationship to the fact table are important for reliable date calculations.
• YTD: Year-to-date accumulation.
• Previous period/year: Compare with an earlier period as requested. MCQ: Time-based comparison/accumulation → use date table and suitable time-intelligence calculation.

Visual calculations

• RUNNINGSUM: Running/cumulative total across the visual's displayed order.
• WINDOW: Refers to a window/range of rows relative to the current row in visual calculations. MCQ traps: Running total → RUNNINGSUM; a range of rows before/after/current row → WINDOW.

Performance Analyzer

• Performance Analyzer helps identify how long report visuals take and inspect their performance/query timing.
• Use it to diagnose a slow report visual.
• It does not primarily diagnose scheduled refresh history; refresh history is used for refresh outcomes/errors. MCQ clue: Find which visual is slow → Performance Analyzer.

MODULE 5 — DESIGN EFFECTIVE REPORTS & ANALYTICS

Visual selection

• Bar/column: Compare categories.
• Line: Trend over time.
• Card/KPI: One important metric/status.
• Table: Exact values/records.
• Matrix: Values grouped by rows and columns, often hierarchical.
• Scatter: Relationship between two numeric variables.
• Map: Geographic pattern when location matters.
• Decomposition tree: Break a metric down by dimensions to explore contributors.
• Key Influencers: Identify factors associated with an outcome; does not prove causation.
• Smart narrative: Text summary of data/visuals.
• Q&A: Ask supported natural-language questions about model data.
• Forecast: Estimate future values in a supported time-series visual.
• Anomaly detection: Flag unusual points in a time series.

Filters and interactions

• Visual-level filter: One visual.
• Page-level filter: Visuals on one page.
• Report-level filter: Across report pages.
• Slicer: On-page interactive filter control.
• Cross-filter: Selection filters another visual.
• Cross-highlight: Highlights related data while leaving other data visible.
• Relative date slicer: Choose relative periods, such as last 30 days.

Drill and navigation

• Drill-down: Move to a lower hierarchy level inside the same visual (Year → Quarter → Month).
• Drill-up: Return to a higher hierarchy level.
• Drillthrough: Open a separate detail page filtered to the selected item.
• Tooltip: Extra information on hover; report-page tooltip can show a custom detail view.
• Bookmark: Saves a report view/state, depending on settings.
• Button: Triggers actions such as navigation, bookmark or drillthrough. MCQ trap: Same visual hierarchy = drill-down; separate filtered detail page = drillthrough.

MODULE 6 — MANAGE, DISTRIBUTE, REFRESH & SECURE CONTENT

Workspace roles

• Admin: Broad workspace control, including access/settings.
• Member: Broad content collaboration and management capabilities.
• Contributor: Creates/edits content but has fewer management rights than Member.
• Viewer: Views content without editing. MCQ clue: Give only necessary access = least privilege.

Workspace, app, report, dashboard

• Workspace: Team collaboration and content management.
• App: Curated package for distributing content to consumers.
• Report: Interactive, potentially multi-page analysis.
• Dashboard: One-page Power BI Service canvas built from pinned tiles.

Sharing and Build permission

• Build permission: Allows creating new reports/content from a semantic model, subject to other access and licensing requirements.
• Workspace role, item sharing, app audience and semantic-model permissions are distinct controls.

Refresh and gateway

• Scheduled refresh: Updates imported data from its source on a schedule.
• Refresh history: Shows refresh outcomes and error details.
• Common failures: expired credentials, source unavailable, changed schema, gateway offline, timeout.
• On-premises data gateway: Secure connection between Power BI Service and a source it cannot directly reach, often on-premises. Needed only when the source/connectivity requires it. MCQ clues: Refresh failed → refresh history. Cloud service must reach on-prem source → gateway.

Endorsement

• Promoted: Recommended/useful content endorsed by its owner or creator.
• Certified: Formally validated under organizational criteria by authorized people/process.

RLS and OLS

• Row-level security (RLS): Limits which rows a user can see, e.g. a manager sees only their region.
• Static RLS: Fixed filter rules/assignments.
• Dynamic RLS: Uses the signed-in user's identity and a mapping table to determine permitted rows.
• USERPRINCIPALNAME() commonly returns the current user's sign-in identity and is often used for dynamic RLS.
• Object-level security (OLS): Restricts access to model objects such as tables or columns. MCQ traps: Rows → RLS; tables/columns → OLS. Hiding a column in a visual is not security. Test security with intended user roles.

Audit and usage

• Audit logs: Investigate recorded activities such as sharing, publishing or access, depending on available configuration.
• Usage metrics: Understand report usage/adoption and view activity where supported.

HIGH-YIELD MCQ MATCHING TABLE

LAST-MINUTE EXAM STRATEGY

1. Identify the task word in the question: clean, join, calculate, filter, visualize, share, refresh or secure.
2. Match the task to the correct tool/function—not just a familiar term.
3. Watch for near-synonyms: Merge ≠ Append; RLS ≠ OLS; drill-down ≠ drillthrough; report ≠ dashboard; query folding ≠ data reduction.
4. If the question asks for the best first step, clarify business requirements before choosing visuals or implementation.
5. For security, distinguish permissions (who can access/build) from RLS/OLS (what data/model objects they can see).
6. For slow visuals, use Performance Analyzer; for failed refreshes, inspect refresh history and credentials/gateway.