Module 4 — Visualize & Analyze Data

Choosing a visual is not mainly about appearance; it is about matching the analytical question to the visual encoding that communicates the answer most accurately. A good report uses the simplest visual that makes the intended comparison, trend, relationship or status easy to understand.

Common visual purposes

• A line chart visual is primarily used to show trends or changes in a measure over a continuous sequence, most commonly time. The connected points emphasize direction and movement from one period to the next.
• A bar or column chart is suited to comparing values across discrete categories such as products, departments or regions. Horizontal bars are especially useful when category labels are long.
• A stacked bar/column chart adds a part-to-whole view within each category, but too many series can make the composition difficult to read.
• A pie or donut chart shows part-to-whole composition for a small number of categories. It becomes difficult to compare accurately when many slices have similar sizes, so bar charts are often clearer for detailed category comparison.
• A scatter chart shows the relationship between two numeric measures by plotting one on the X-axis and another on the Y-axis. Size and legend can add extra analytical dimensions.
• A card emphasizes one or a small number of important values. A KPI visual is designed around status and trend relative to a target, while a gauge shows a current value against a target or range.
• A table is best for detailed records or precise values. A matrix adds row/column groupings, hierarchical drill behavior and cross-tab style analysis.
• Map visuals should use genuine geographic fields and suitable data categories such as City, State/Province, Country/Region, Postal Code, Latitude or Longitude.

Configure the visual for the question

After selecting a visual type, configure axes, legends, sort order, data labels, titles and aggregation deliberately. For example, a bar chart sorted by descending sales answers “which categories are largest?” more directly than alphabetical sorting. A time-series line chart should normally use a chronological date field rather than a text month name that sorts alphabetically.

Power BI reports are interactive because filters can be applied at different scopes and visual selections can affect other visuals. Understanding where a filter applies and how interactions propagate is essential for predictable report behavior.

Filter scopes and slicers

• A visual-level filter affects only one visual.
• A page-level filter affects all applicable visuals on the current report page.
• A report-level filter affects applicable visuals across the report.
• A slicer is an on-canvas filtering visual that lets the report consumer change filter context interactively.

Top N filtering does more than sort. You specify how many categories to keep and provide a numeric measure used to rank them. Sorting a visual in descending order changes display order but does not remove the remaining categories.

Edit interactions

When a user selects data in one visual, Power BI can make another visual respond in three main ways. Filter restricts the target visual to the selected subset. Highlight preserves the target's broader context but visually emphasizes the selected portion. None prevents that source visual from affecting the target.

Sync slicers and parameter-driven interaction

Sync slicers lets one slicer selection apply to multiple report pages. A synchronized slicer can be hidden on a page while its filter still applies there.

Field parameters can let consumers switch which dimension or measure a visual uses—for example, changing an axis between Region, Product and Salesperson. What-if parameters are different: they create a numeric input that supports scenario analysis, such as changing an assumed growth rate.

Bookmarks, buttons, tooltips and drill features help a report behave more like an analytical application rather than a collection of static charts. The key is to understand which state is being saved or passed when the user navigates.

Bookmarks and the Selection pane

A bookmark stores a report state. Depending on its settings, it can capture Data state such as slicer/filter selections, Display state such as which objects are visible, and the Current page. The Selection pane lets you control visibility and layer order for report objects.

If a bookmark is intended only to show and hide a help panel, menu or alternate visual, avoid capturing Data unless you deliberately want the bookmark to restore slicers and filters. Otherwise, clicking the bookmark can unexpectedly reset the user's analytical selections.

Tooltips, drill down and drillthrough

• A report page tooltip is a specially configured page that appears when a user hovers over a data point. It can contain multiple visuals and gives extra context without navigating away.
• Drill down moves to a deeper level of a hierarchy within the same visual—for example, Year → Quarter → Month.
• Drillthrough navigates to another report page while carrying a selected value as filter context. It is appropriate for moving from summary to detail, such as right-clicking a customer and opening a Customer Details page.
• Navigation buttons can link to pages, bookmarks, drillthrough destinations and other supported targets, helping create a guided report experience.

Drillthrough can also be configured to keep additional filters so the destination page receives more of the user's current context, not just the primary drillthrough field.

A technically correct report can still fail if it is difficult to read, inaccessible, inconsistent or awkward on the user's device. Report design therefore includes styling, accessibility, page configuration, mobile layout and controlled personalization.

Themes and conditional formatting

A theme applies consistent visual formatting across a report, including colors, fonts and default visual properties. Themes can be customized using JSON, which is useful when an organization wants repeatable branding across many reports.

Conditional formatting changes visual appearance based on data. Rule-based formatting uses explicit thresholds; color scales create a continuous gradient; field-value formatting allows a field or measure to return a formatting value such as a color. Icons and data bars can provide additional visual cues.

Accessibility

• Add meaningful alt text so screen-reader users can understand the purpose of important visuals.
• Configure a logical tab order for keyboard navigation.
• Maintain sufficient color contrast and readable text sizes.
• Do not rely on color alone to communicate meaning; combine color with text, icons, labels or position.

Mobile, page settings and export

Mobile layout lets you arrange visuals specifically for a phone screen without redesigning the desktop report. Report page settings also control properties such as page size, wallpaper/background and—in supported scenarios—automatic page refresh, which periodically updates a page so DirectQuery monitoring reports can show fresher results.

Report authors can configure export settings to control whether consumers may export summarized or underlying data where supported. This is a usability/governance choice, not a replacement for security such as RLS.

Personalize visuals

When personalization is enabled, consumers can change fields, visual types or other properties for their own exploration without editing the published report for everyone else. This supports self-service analysis while preserving the original report design.

Power BI contains analytical features that help users move beyond displaying values and begin identifying patterns, unusual behavior, uncertainty and possible future movement. Each feature answers a different type of question.

Reference and analytical lines

A constant/reference line displays a fixed benchmark such as a target of 95%. An average line is calculated from the data shown in the visual. Other supported analytical lines can represent minimum, maximum, median or percentile-type reference information depending on the visual.

Forecasting, anomalies and error bars

• Forecasting uses historical time-series patterns to estimate future values and can display an uncertainty interval. It is appropriate when the question is “what might happen next?” rather than “which historical point is unusual?”
• Anomaly detection identifies unexpected spikes or dips in supported line-chart scenarios. It is intended to highlight observations that differ from the expected pattern in existing data.
• Error bars communicate uncertainty or variation around a value. They can help viewers understand that a point estimate has a range rather than treating it as perfectly precise.

Grouping, binning and clustering

Grouping manually combines known discrete categories, such as placing several small product categories into “Other.” Binning places numeric or date values into ranges such as Age 20–29 or Sales 0–999. Clustering is different because Power BI uses patterns in numeric data to identify naturally similar groups of observations.

Analyze feature

The Analyze functionality can help explain increases, decreases or differences in a visual by examining other fields that may contribute to the observed change. It is an exploratory aid, not proof of causation.

Power BI's AI-assisted features help users investigate drivers, decompose measures, ask natural-language questions and generate narratives or report content. These features overlap in purpose, so exam questions often test whether you can match the feature to the analytical task.

Key influencers

The Key influencers visual analyzes a target outcome and ranks the factors that are most strongly associated with that outcome. The target might be categorical—such as whether a customer churned—or numeric. The visual helps answer questions such as “which attributes are associated with higher satisfaction?” or “what factors are associated with churn?”

Key influencers identifies statistical associations in the available data; it should not automatically be interpreted as proving that an influencer caused the outcome.

Decomposition tree

A decomposition tree starts with a measure and lets the user break it down across multiple explanatory dimensions. Users can choose the next split manually or use AI-assisted high/low splits. It is ideal for interactive root-cause-style exploration such as Total Profit → Region → Product Category → Sales Channel.

Q&A, narratives and Copilot

• Q&A allows the user to type a natural-language question such as “sales by region this year” and Power BI generates an appropriate result or visual based on the semantic model.
• Narrative visuals generate textual descriptions of important report insights. In current Power BI experiences, Copilot can assist with narrative summarization and report creation where the required tenant/capacity features are available.
• Copilot can help suggest content for a new report page, create report content from prompts, and summarize the underlying semantic model. Good model names, descriptions and business terminology improve the quality of AI-assisted experiences.

Standard Power BI reports are optimized for interactive exploration, but some requirements are primarily about precise printed layout, complete multi-page output or calculations that depend on the structure of a visual. Paginated reports and visual calculations address these different needs.

Paginated reports

A paginated report is designed for highly formatted, page-oriented output. It uses the RDL report format and is commonly authored with Power BI Report Builder. Unlike a normal interactive report canvas, a paginated report is built with explicit control over page size, headers, footers, tables, grouping, page breaks and print layout.

Paginated reports are appropriate for scenarios such as invoices, statements, regulatory extracts, operational lists and reports where users must export or print every row across many pages with repeating headers. The defining requirement is not simply “the dataset has many rows”; it is that the output needs pixel-precise, print-friendly pagination.

Visual calculations

A visual calculation is a DAX calculation evaluated over the data structure produced by a visual. This makes it convenient for calculations such as running totals, moving averages or comparisons that depend on the visual's rows and columns.

Because visual calculations belong to the visual layer, they are not the same as reusable semantic-model measures. If a calculation represents governed business logic that should behave consistently across many reports and visuals, a model measure—or sometimes a calculation group—is usually the better design.