Power BI Desktop
The free Windows application where you connect to data, clean it, model it, write calculations and design reports. The whole of this course except sharing happens here.
Sixty-eight terms from the course, defined for development practitioners rather than corporate analysts. Search, or filter by part of the workflow.
The free Windows application where you connect to data, clean it, model it, write calculations and design reports. The whole of this course except sharing happens here.
The cloud at app.powerbi.com where finished reports are published and shared. Viewing and light editing work in any browser; publishing needs a paid licence.
The phone and tablet apps for viewing published reports. A consumption surface, not an authoring one.
The Power BI file format. A single file that can bundle the queries, the data model with its calculations, and the report pages, which makes a whole analysis portable.
The canvas where you build and arrange visuals. The first of the three view icons on the left edge.
The grid that shows the actual rows of a loaded table, for inspecting values. The middle view icon.
The diagram where tables and their relationships are shown and edited. The third view icon, where you build a star schema.
The command bar across the top: Home, Insert, Modeling, View and others. Get Data and Refresh live on Home.
The list on the right of every table, column and measure in your model. You drag from here onto visuals.
The panel where you choose a chart type and drag fields into wells such as Axis, Values and Legend.
The Home-ribbon entry point to over a hundred connectors: Excel, Text/CSV, Web, databases and more. Where every project begins.
A built-in method for reading a particular kind of source, such as the Excel, CSV or Web connector.
The button that brings data straight into the model as-is, skipping cleaning. Usually the wrong default.
The button that opens the Power Query Editor so you can clean before loading. Usually the right default.
The environment for cleaning and reshaping data through a recorded sequence of steps.
The ordered list of every transformation applied to a query. An audit trail and a recipe that re-runs on each refresh.
The functional language behind Power Query. Each Applied Step is M code; the Advanced Editor shows it.
Turn the first row of values into column names. Common after removing a stray title row from a government export.
Set a column's data type, such as decimal number or text. Wrong types are the most common cause of broken totals.
Drop unwanted rows in Power Query, such as blanks, footnotes or a Total row that would double-count.
Transformations that remove leading and trailing spaces (Trim) and non-printing characters (Clean), fixing keys that look equal but are not.
Swap one value for another across a column, such as turning the text NA into a true null.
Break one column into several by a delimiter, such as separating State and District from a combined label.
Reshape a wide table (one column per indicator) into a long one (one row per observation). The key move for survey factsheets.
Unpivot everything except a chosen key column, so newly added indicators flow in automatically on future files.
Join two queries by a matching key, bringing columns from one into the other. Supports left, inner, full outer and anti-joins.
Stack the rows of two queries into one, a union. Used to combine survey rounds into one long table.
Reference creates a query that points at another and follows its changes; Duplicate makes an independent copy.
The command that exits Power Query and loads the cleaned tables into the model.
A model with one central fact table joined to surrounding dimension tables. The shape to aim for in nearly all cases.
The table of events or measurements, one row per observation, holding the numbers you aggregate and keys to the dimensions.
A table of descriptive context, each item listed once, such as a District or Date table. What you slice and group by.
A defined link between two tables on a shared key, letting filters and calculations flow between them.
The nature of a relationship: usually one-to-many, from a dimension to the fact table.
Which way a filter flows across a relationship. Single is the safe default; bidirectional only for a named need.
A dedicated table with one continuous row per period, marked as a date table, required for reliable time intelligence.
The column used to match rows across tables in a relationship, such as a district code. Must share type and be clean.
Data Analysis Expressions, the formula language for calculations. Looks like Excel formulas; behaves differently.
A calculation evaluated on demand against the current filter context, not stored. The right tool for rates, totals and shares.
A calculation evaluated row by row at refresh and stored in the table. Used sparingly, for things like a classification flag.
The single-row perspective in which a calculated column is evaluated.
The set of filters active when a measure is evaluated, set by slicers, the visual, and any CALCULATE around it.
Basic aggregate functions over a column.
Safe division that returns blank instead of an error when the denominator is zero. Use it instead of the slash, always.
The central DAX function: it evaluates a measure while modifying the filter context, which is how comparisons are built.
Returns a filtered table, often used inside CALCULATE for conditions a simple equals cannot express.
Counts the unique values in a column, such as how many distinct districts reported.
Variables in a DAX expression, used to make multi-step measures readable and to compute a value once.
Calculations comparing periods, such as change between survey rounds, which depend on a proper date table.
Encodes magnitude as length, read accurately. The default for comparing categories; sort it for an instant ranking.
Encodes change over an ordered axis such as time. Not for unordered categories.
A map shading regions by a value. Map a rate, not a raw count, or it mostly shows where people live.
A visual showing one big headline number, such as a national rate.
Visuals for exact values across one or two dimensions, when the reader needs numbers rather than a shape.
An on-canvas control that filters the page, such as a state selector.
A value axis that does not start at zero, which exaggerates differences. Avoid for bar charts.
Colouring values by a rule, such as turning districts above a threshold amber. Updates as data refreshes.
A written description of a visual for screen readers and accessibility, part of responsible report design.
Clicking a data point in one visual filters the others on the page. On by default; tunable via Edit interactions.
A right-click jump to a dedicated detail page that carries the selected item's filter.
A saved page state (filters and visibility) that a button can trigger, used to build view toggles.
A mini-page set as a hover tooltip, showing custom detail instead of a bare value.
A small file that sets palette and fonts across a whole report, the Power BI equivalent of design tokens.
A free option that creates a fully public, unauthenticated link. For open content only, never for personal or sensitive data.
The per-user licence (14 US dollars per month in 2026) needed to publish and to view shared reports outside large capacity.
Roles that restrict which rows a user sees. Defined in Desktop but only enforced once published with a paid licence.
Free static export from Desktop, suitable for a board pack or email attachment.
India's Digital Personal Data Protection Act, governing digital personal data. Favour aggregates and keep identifiable records off shared artifacts.
No terms match. Try a broader search or a different filter.