Microsoft Power BI Glossary from A to Z — Microsoft Power BI

Microsoft Power BI Glossary — 128 definitions

All the terms defined in the 26 sections of the guide, listed in alphabetical order.
Tip: press Ctrl + F to search for a term on this page.

Scheduled refresh: Semantic model settings → Refresh: the times (up to 8/day on Pro, 48 on Premium) at which the Service reloads the data.
View as: The test in Desktop (Modeling → View as): seeing the report through a role's eyes before publishing.
Implicit aggregation: A numeric field dropped into a value is aggregated automatically (sum by default) — changeable via the field menu.
Align / Distribute: Format → Align: computed edges and spacing — never align by eye.
ALL: Removes filters: ALL(Sales[City]) ignores the city filter — the key to “% of total”: DIVIDE([Revenue]; CALCULATE([Revenue]; ALL(Sales[City]))).
Performance Analyzer: View → Performance Analyzer: start recording, refresh the visuals — each visual shows its detailed timings.
App: The workspace published as a clean package: navigation, differentiated audiences, controlled updates — the wide-distribution channel.
Button + action: Insert → Buttons: each button carries an action (bookmark, page navigation, back) — testable with Ctrl + click in Desktop.
Back button: Automatically added on the drill-through page: returns to the source page (Ctrl + click in Desktop).
Business Intelligence (BI): Turning data into decision-ready information: collect, model, visualize, distribute.
CALCULATE: THE DAX function: evaluates an expression in a modified filter context — previewed here, mastered in level 3.
CALENDARAUTO / CALENDAR: The DAX functions that generate the calendar: CALENDARAUTO() sweeps the model's dates; CALENDAR(start; end) sets them explicitly.
One-to-many cardinality (1:*): The standard case: ONE product from the repository ↔ MANY sales rows. The “1” side is the dimension, the “*” side the facts.
Card (KPI): The “big number” visual: total, average — the opening of every dashboard.
Drill-through field: The field dropped into the target page's “Drill through” well: it defines what can be drilled into.
Load vs Transform: Load = import as-is; Transform = open Power Query first. Professional reflex: always Transform.
Calculated column: New DAX column evaluated row by row and STORED: Margin = Sales[Amount] - Sales[Cost]. Bloats the model.
Key column: The shared column carrying the relationship: unique on the dimension side, repeated on the facts side — never duplicates on the “1” side.
Work or school account: The Service requires a work/school account (no personal accounts such as Gmail): your organization's account.
Folder connector: Points to a folder of uniform files: Power BI combines them into a single table — the anti-copy-paste for monthly exports.
Live connection: The report consumes a published model as-is: one single source of truth, several connected reports.
Filter context: A measure's environment: “what the visual filters” (the bar's city, the column's month) — covered in depth in 3.1.
Row context: A calculated column's environment: “the current row” — each reference points to that row's value.
Copy query: Each slow measure is copied (Copy query) for analysis in DAX Studio — the practitioners' free tool.
Accent color: THE color carrying the key information; everything else in neutrals — the anti-rainbow rule.
COUNTROWS vs COUNT: COUNTROWS(Sales) counts a table's rows (the most useful); COUNT counts the non-blank values of a column.
Dataflow: Power Query in the cloud: centralized queries in the Service, reusable across reports — prepare once, consume everywhere.
DATEADD: The generic offset: DATEADD(Dates[Date]; -1; MONTH) for the previous month — more flexible than dedicated functions.
DAX: The language of measures and calculated columns: Excel-like functions, but evaluated in a model context.
Unpivot: Turning crossed columns (Jan, Feb, Mar…) into two Attribute/Value columns: THE transformation that makes data analyzable.
Automatic detection: Power BI guesses some relationships on import: ALWAYS verify in Model view — it is sometimes wrong.
DirectQuery: Data stays in the source (live queries): for very large volumes or real time, at the cost of performance.
Mobile layout: View → Mobile layout (Desktop): rearranging existing visuals vertically for phones.
DIVIDE: DIVIDE(numerator; denominator; [if_zero]): bulletproof division — never a divide-by-zero error again.
Drill-through: The detail page reached via right-click → Drill through: it arrives automatically filtered on the source item.
Power Query Editor: The preparation workshop (Transform data): codeless transformations, recorded as steps.
Workspace: The Service container: reports, models, dashboards of the same scope — with roles (admin, member, contributor, viewer).
Applied steps: The journal on the right: each transformation listed, editable, deletable — the recipe replayed at each refresh.
Year-over-year change: The final duo: Change = [Revenue] - [Revenue PY] and Change % = DIVIDE([Revenue] - [Revenue PY]; [Revenue PY]).
Official practice assessment: The Microsoft mock test (Practice Assessment): free, under real conditions — the thermometer of your preparation.
Cross-filtering: Clicking an item in one visual filters the other visuals on the page: Power BI's native interactivity.
Merge / Append: Merge = join between two queries (like XLOOKUP); Append = stacking tables with the same structure.
Theme generator: Community tools (e.g. themes.powerbi.tips) that produce the JSON from your brand colors.
Granularity: A fact row's level of detail (the sale? the order line? the day?): to be defined consciously, never mixed.
Hierarchy: Year → Quarter → Month → Day: drilling through levels in visuals (drill buttons).
Refresh history: The log of attempts (successes/failures + detailed errors): the first place to look when “data is not up to date”.
Model hygiene: Delete unused columns (in Power Query, not after), reduce cardinalities (no useless unique IDs), disable auto date/time.
Custom tooltip: A dedicated page (Page information → Tooltip) shown on hover: contextual mini-visuals on the hovered item.
Credentials: The source credentials the Service uses to refresh: configure them in the model settings — the #1 cause of refresh failures.
Microsoft Store installation: The recommended route: automatic monthly updates (Power BI evolves every month — seriously).
Interactions: The cross-behavior of visuals: Format → Edit interactions → for each pair, filter / highlight / none.
The 4 domains: Prepare data (~25-30%), model (~25-30%), visualize/analyze (~25-30%), manage and secure (~15-20%) — indicative weights, check the official page.
Viewer licensing: On Pro capacity: each reader needs a Pro license; on Premium/Fabric: readers consume without a Pro license.
Free license: Create in Desktop + publish to YOUR personal workspace: enough to learn, not enough to share.
Pro license: The business standard (per user/month): sharing, collaborative workspaces, apps.
Mark as date table: The official setting (table tools) enabling the correct time-intelligence behavior.
Matrix: The pivot-table equivalent: rows, columns, values, subtotals and hierarchical drill-down.
Measure: DAX calculation evaluated on demand in the visual's context: Total Revenue = SUM(Sales[Amount]). Stores nothing.
Implicit measure: The automatic aggregation of a column dropped into values: handy at first, to be replaced by real explicit measures.
Microsoft Learn: The official FREE preparation path: modules aligned with the exam — the backbone of your preparation.
Conditional formatting: Data-driven color (Format → color → fx): red appearing ONLY if margin is negative.
Import mode: Data is copied and compressed into the .pbix file: maximum performance — the default, recommended mode.
Composite model: The mix: connection to a published model + added local tables (DirectQuery on semantic model) — maximum flexibility, proportional complexity.
Semantic model (dataset): The published model (tables, relationships, measures): other reports can connect to it live without duplicating.
Published semantic model: Data and measures in the Service: reusable, refreshable, the basis of live connections (4.5).
My workspace: The personal workspace: perfect for learning, forbidden for team production (everything depends on one person).
Page/bookmark navigator: Auto-generated buttons (Insert → Buttons → Navigator): the effortless navigation bar.
Get data: The universal entry point (Home): 200+ connectors, from files to databases to the web.
Bookmark options: A bookmark can capture Data (filters) and/or Display (visibility): the fine-tuning that avoids surprises.
Field parameter: Modeling → New parameter → Fields: the list of measures/columns a slicer can swap into visuals.
Power Query parameter: A query parameter (file path, environment): the source becomes configurable without rewriting steps.
What-if parameter (numeric): Modeling → New parameter → Numeric range: a slider whose value (SELECTEDVALUE) is used in measures.
Direct sharing: A report's Share button: quick, but unmanageable beyond a few people — troubleshooting, not architecture.
Gateway: The bridge software installed on a company server: lets the Service reach on-premises sources securely.
Personal vs standard gateway: Personal: on your PC, for you (depends on your machine being on!); standard/enterprise: on a server, shared, supervised by IT — the only serious option.
Customize theme: The built-in editor (Themes → Customize): colors, text, backgrounds — without touching the JSON.
PL-300: Microsoft's official “Power BI Data Analyst” exam: MCQs, case studies, sometimes hands-on — in a test center or remotely proctored.
Power BI Desktop: The FREE Windows creation app: data connections, model, DAX measures, reports. This whole guide starts there.
Power BI Service: The cloud platform (app.powerbi.com): publishing, workspaces, sharing, scheduled refresh, dashboards.
Premium / Fabric: Dedicated capacity for large volumes and wide distribution (consumers no longer need Pro), integrated into the Microsoft Fabric ecosystem.
Publish: Home → Publish (Desktop): the .pbix goes up to the Service, split into report + semantic model.
Field wells: A visual's zones (X/Y axis, legend, values, tooltips) where columns and measures are dropped.
User-driven report: The result: the user picks the angle, the measure, the scenario — the same report serves ten analyses.
Report vs dashboard: The report = pages of interactive visuals created in Desktop; the dashboard (Service) = a compilation of tiles pinned from one or more reports.
Relationship: The link between two tables on a shared column (Products[Code] ↔ Sales[ProductCode]): the channel through which filters flow.
Renewal: The certification renews each year via a FREE online assessment: a differentiator that stays current.
DAX query vs Rendering: The two timed items: slow query = model/measure to review; slow rendering = overloaded visual.
RLS (Row-Level Security): Filtering visible ROWS by signed-in user: one report, different data scopes.
Role: A named set of DAX filters (Modeling → Manage roles): “Lyon Sales” = Sales[City]="Lyon".
Workspace roles: Admin, member, contributor (produce) vs Viewer (consumes): structured team access.
SAMEPERIODLASTYEAR: The same period one year earlier: CALCULATE([Total Revenue]; SAMEPERIODLASTYEAR(Dates[Date])).
Star schema: Facts at the center, dimensions around, 1:* relationships: Power BI models' reference architecture.
Slicer: The interactive filter placed ON the page: list, dropdown, date range, buttons — the report's user interface.
SELECTEDVALUE: The function reading the parameter's chosen value: Simulation = [Total Revenue] * (1 + [Increase Value %]).
Filter direction: By default, filters flow from the “1” side to the “*” side: filtering a product filters its sales. Single direction is the healthy norm.
Bookmark: A snapshot of the report state (View → Bookmarks → Add): filters, selections, visibility — recallable at will.
SUM / AVERAGE / MIN / MAX: Basic aggregations on a column: SUM(Sales[Amount]).
Cross-highlighting: The default mode of some visuals: the selected part is highlighted instead of filtering — switchable to “filter” if needed.
Sync slicers: View → Sync slicers: the same slicer acts across pages — navigation consistency.
Date table: Calendar dimension: one row per day, continuous over the period, linked to the facts' date — the foundation of time analysis.
Dimension table: An analysis axis: one row = one unique item (a product, a customer, a day), with descriptive attributes.
Fact table: Events to measure: one row = one transaction/event, with keys to dimensions and numeric values.
Security table: The email → scope (city, branch…) mapping table: a single dynamic role flows through it to filter.
Single table (anti-pattern): Everything in one big table: it “works” at first, then DAX and performance degrade — the star schema is the remedy.
Dashboard: The Service's tiles page: key visuals pinned from one or more reports — the executive view.
Page size: 16:9 by default; customizable (Page format) per target: screen, portrait for mobile/PDF.
Tooltip page size: The predefined compact format (Page format → Size → Tooltip): a tooltip must stay a tooltip.
Report theme (JSON file): The JSON file (View → Themes): palette, fonts, per-visual default settings — applied to the whole report at once.
Report theme (palette): View → Themes: the global color and font palette (detailed in level 4.1).
Time intelligence: The DAX function family handling periods: totals, offsets, comparisons — all backed by the date table.
TOTALYTD: Year-to-date total: TOTALYTD([Total Revenue]; Dates[Date]) — fiscal variant possible with the year-end date.
Context transition: The (advanced) mechanism by which CALCULATE turns row context into filter context — worth knowing by name, worth digging into with SQLBI.
Sort by column: The month “April” sorts before “January” alphabetically: sort the Month column by the MonthNumber column (Column tools → Sort by column).
Too many visuals: Each visual = one query: a page with 15 visuals sends 15 queries on every click. Sobriety is an optimization.
Data type: Each column must be correctly typed (text, whole/decimal number, date): half of beginners' errors come from there.
One single source of truth: The architecture principle: key measures defined ONCE, consumed by all — the end of “different revenue per report”.
USERPRINCIPALNAME(): The function returning the signed-in user's email: the key to dynamic RLS.
VAR / RETURN: DAX variables: VAR TotalOverall = … RETURN … — readable formulas, no repeated calculations, easier debugging.
Data (Fields) pane: All available tables and columns: the raw material dragged into visuals.
Filters pane: Author filters: on THIS visual, on THIS page, or on the WHOLE report — visible or hidden to readers per setting.
Format (paint roller) pane: Formatting of the selected visual: title, colors, labels, background, border, shadow.
Selection pane: The page's object list with visibility (eye): the basis of show/hide toggles.
Visualizations pane: The visual-type picker and its field wells (axis, values, legend) + formatting (paint roller).
Model view: The big picture: tables, relationships, cardinalities — level 2's architecture.
Report view: The page canvas where visuals are built: your main workplace.
Table (Data) view: Reading imported tables row by row: checking, sorting, creating columns — without directly editing cells.
BI workflow: The full cycle: connect → clean (Power Query) → model (relationships, DAX) → visualize → publish → refresh.
Canvas area: The page space: what overflows doesn't exist — the constraint forcing prioritization.