Selecting and Configuring Core Visuals

Introduction

Power BI visuals convert measures, dimensions, and business rules into forms that support comparison, trend analysis, composition analysis, distribution analysis, and decision-making. Selecting a visual is therefore not merely a formatting decision. It is an analytical design decision: the visual must match the question being asked, represent the data accurately, and allow users to identify the intended message quickly.

A well-designed report distinguishes between the data model and its visual presentation. The model supplies fields, relationships, calculated columns, and measures; the visual determines how those results are perceived. A technically correct measure can still mislead if it is displayed in an inappropriate chart, uses an unsuitable aggregation, or is surrounded by excessive decoration.

This chapter explains the principal Power BI visuals and the circumstances in which each should be used.

Learning Objectives

By the end of this chapter, you should be able to:

• match common business questions with suitable Power BI visuals;
• configure tables, matrices, cards, KPIs, and gauges;
• create and format bar, column, line, area, and combo charts;
• use scatterplots, treemaps, waterfall charts, funnels, maps, and other core visuals;
• apply appropriate aggregation, sorting, filtering, labelling, and interaction settings;
• evaluate visual limitations, accessibility, and governance risks.

Analytical Purpose and Visual Selection

Before choosing a visual, identify four characteristics of the question:

1. Comparison: Which category is larger or smaller?
2. Trend: How does a value change over time?
3. Composition: How is a total divided among parts?
4. Relationship or distribution: Are variables associated, or how are values spread?

A fifth consideration is precision. If users must read exact values, a table or card may be preferable to a chart. If users must identify a pattern rapidly, a chart is usually more effective.

A useful design principle is to place the most important visual in the most prominent position and to avoid presenting several visuals that communicate the same fact without adding insight.

Tables, Matrices, Cards, and KPIs

Tables

A table displays records or summarized values in rows and columns. It is appropriate when users require detailed inspection, exact figures, or exportable information. Typical fields include customer, product, date, quantity, revenue, and margin.

In Power BI, configure a table by placing categorical fields and measures in the Columns well. Measures should generally be used for business calculations because they respond to filters and slicers. Important configuration options include:

• sorting by a meaningful measure rather than alphabetically;
• applying conditional formatting to values, background, or data bars;
• choosing appropriate decimal places and currency symbols;
• hiding unnecessary technical columns;
• using tooltips for secondary information;
• limiting the number of rows where a full detail listing would overwhelm the report.

Conditional formatting should reinforce meaning. For example, red and green may indicate performance against a target, but the colour should not be the only indication because colour-blind users may not distinguish it.

Matrices

A matrix is a multidimensional table. It supports row groups, column groups, subtotals, grand totals, and drill-down. It is appropriate for hierarchical analysis such as Region → Country → City or Year → Quarter → Month.

A common configuration places a hierarchy in Rows, a second dimension in Columns, and one or more measures in Values. Designers should decide whether subtotals are analytically useful. Excessive subtotals can make a matrix difficult to scan, while omitting them may remove important context.

A matrix is not automatically a pivot table in the accounting sense; its values still depend on the model, filter context, and measure definitions. Drill-down should be enabled only when the hierarchy is meaningful and consistently ordered.

Cards

A card displays one prominent value, such as total revenue, orders, customers, or current inventory. Cards are useful for executive summaries and headline metrics, but they provide little context by themselves. A card showing “£2.4 million” is more informative when accompanied by a period label, comparison, or target.

Use a measure rather than a raw numeric column when the value must respond correctly to report filters. Configure display units, decimal precision, category labels, and title text carefully. A card should not contain so many values that it becomes a small, poorly organised table.

KPI visuals and gauges

A KPI visual compares an indicator with a target and often displays a trend. A typical KPI consists of an actual measure, a target measure, and a time or category axis. For example, actual monthly sales may be compared with a budget measure.

A gauge communicates progress toward a goal. It is suitable when a single value must be compared with a clearly defined minimum, target, or maximum. Gauges are less suitable for comparing many categories because each gauge requires separate interpretation.

Targets must be defined precisely. A target may be a fixed value, a budget measure, a prior-period value, or a statistically calculated benchmark. Users should be told which interpretation is used. A favourable result is not always “higher”: for costs, defects, or response time, lower may be better.

Bar and Column Charts

Bar and column charts encode values using length. A bar chart uses horizontal bars; a column chart uses vertical bars. Both are effective for comparison across discrete categories.

Use a bar chart when category names are long, when there are many categories, or when ranking is important. Use a column chart when there are relatively few categories or when the horizontal axis represents time periods. Sort bars by the measure when the purpose is ranking; retain natural or chronological order when sequence matters.

Important configurations include:

• axis type: categorical or continuous;
• clustered versus stacked arrangement;
• data labels and display units;
• small multiples for repeated comparisons;
• reference lines for averages, targets, or thresholds;
• drill-through or report-page tooltips for detail.

Clustered bars compare separate series. Stacked bars show both total and composition, but comparisons of interior segments are difficult because they do not share a common baseline. A 100% stacked chart compares proportions rather than absolute totals and should not be used when total volume is important.

Avoid three-dimensional effects, unnecessary gridlines, and truncated axes that exaggerate differences. For bar and column charts, a zero baseline is normally appropriate when bar length represents magnitude.

Line and Area Charts

A line chart is designed primarily for trends across an ordered axis, especially time. The horizontal axis should generally be a genuine date field or a logically ordered period field. Using text month names without a month-number sort column can produce alphabetical order rather than chronological order.

Line charts can display one or several measures. Too many lines create visual competition; small multiples or separate visuals may be more effective. Markers, data labels, and annotations should be used selectively.

An area chart uses filled regions beneath lines. It can emphasise cumulative magnitude and broad movement. However, overlapping areas obscure individual series. A stacked area chart is useful for showing how components contribute to a changing total, but it makes the comparison of non-baseline components difficult.

A moving average can sometimes clarify a trend, but it must be defined and labelled. For a simple moving average over k periods, the value at period t is:

Here, x_t is the observed value at period t and k is the number of periods. This calculation assumes equally spaced observations and a consistent window. It smooths short-term variation but may conceal sudden changes and is unavailable for the first k−1 periods unless another convention is adopted.

Combo Charts

A combo chart combines a column chart with a line chart. It is useful when two related measures have different visual roles, such as sales volume as columns and profit margin as a line.

A secondary axis may be necessary when measures use different scales, but it introduces a significant risk: users may infer a relationship from the visual alignment even when the measures have different units or weak analytical connection. Axis titles, units, and clearly differentiated colours are essential.

A combo chart should not be used merely to place two unrelated measures in one space. If the measures answer different questions, separate visuals may communicate more honestly.

Scatterplots

A scatterplot displays the relationship between two numeric variables. In Power BI, place one measure on the X-axis and another on the Y-axis; use a categorical field for the legend or details and a measure for bubble size when appropriate.

For example, each point might represent a product, with advertising expenditure on the X-axis and revenue on the Y-axis. A trend line may indicate an association, but association does not prove causation. Outliers should be investigated rather than automatically removed. Logarithmic axes can be useful for highly skewed values, but they must be labelled and explained.

Scatterplots are especially valuable for segmentation. Clusters may represent groups of customers or products with similar characteristics. Bubble size adds a third quantitative variable, but too many dimensions reduce interpretability.

Treemaps and Composition Visuals

A treemap represents hierarchical part-to-whole relationships through nested rectangles. Rectangle area represents a measure, while colour can represent a second measure or category. Treemaps work well when there are many categories and limited space.

They are less effective when users must compare similarly sized rectangles precisely. A sorted bar chart is usually better for exact ranking. Treemaps should have clear titles and labels, and colour should not encode a measure whose interpretation conflicts with area.

Pie and donut charts are appropriate only for a small number of mutually exclusive categories that form a meaningful whole. Human perception compares lengths more accurately than angles, so bar charts are often preferable. A category should not be shown as a percentage of a total unless the denominator is clearly defined.

Waterfall Charts

A waterfall chart explains how sequential positive and negative contributions transform a starting value into an ending value. It is useful for financial variance analysis, such as:

Opening profit → price effect → volume effect → cost effect → closing profit.

The beginning and ending bars represent totals; intermediate bars represent increases or decreases. Correct classification of each component is essential. A cost increase may reduce profit, while a cost reduction may increase it. Totals should be labelled clearly, and the chosen sequence should reflect the business logic rather than arbitrary sorting.

Funnel Charts

A funnel chart shows quantities across ordered stages, such as visitors, leads, qualified leads, proposals, and won deals. It is appropriate when each stage is logically part of a progression.

A funnel can reveal stage volume and apparent drop-off. If stage conversion is important, calculate it explicitly:

The formula assumes that stage B is a valid subsequent stage and that the counts use the same population definition and time period. A lower count at a later stage may reflect time delays, incomplete data, or different counting rules rather than process failure.

Maps and Other Core Visuals

Maps are useful when location is analytically relevant. Power BI may use geographic categories such as country, postcode, latitude, and longitude. A geographic field should be assigned the correct data category, and ambiguous names should be disambiguated with additional fields. Use maps to analyse spatial patterns, not merely to decorate a report.

A filled map is suitable for area-based comparison, while point maps are suitable for locations. Large populations or overlapping points may require aggregation or small multiples. Maps can create privacy risks when locations identify individuals or vulnerable groups.

Slicers provide interactive filtering and are particularly useful for date, region, product, or department selection. They are not analytical visuals by themselves, but they shape the filter context of other visuals. Tooltips, drill-through pages, decomposition trees, and key influencers can extend analysis when users need explanation rather than only description.

Configuring a Visual: A Practical Workflow

A reliable workflow is:

1. State the business question in one sentence.
2. Identify the grain of analysis: transaction, customer, product, month, or another unit.
3. Select the measure and verify its aggregation.
4. Choose a visual whose visual encoding matches the question.
5. Add dimensions, hierarchies, legends, and filters.
6. Sort and format the visual for correct interpretation.
7. Add a meaningful title, units, and explanatory tooltip.
8. Test cross-filtering, drill-down, and interactions.
9. Check accessibility, performance, and mobile layout.
10. Ask a colleague to interpret the visual without verbal explanation.

Aggregation deserves particular care. Sum is appropriate for additive quantities such as units sold; averages may be appropriate for prices or response times; distinct count is appropriate for customers when repeated transactions must not be counted as separate customers. A measure such as average margin should not be replaced casually by an average of already aggregated subgroup averages.

Illustrative Case: Regional Retail Performance

Background. A fictional retailer wants a Power BI report for monthly management meetings. The report contains sales transactions, product categories, regions, sales targets, and customer information.

Problem. Managers need to know whether sales are growing, which regions require attention, what explains changes in profit, and where the sales pipeline is weakening.

Analysis and approach. The report designer uses:

• cards for total sales, gross profit, and distinct customers;
• a line chart for monthly sales, with a chronological date axis;
• a sorted bar chart for regional sales comparison;
• a matrix showing region by category with drill-down to products;
• a KPI comparing sales with the monthly target;
• a waterfall chart explaining the difference between budgeted and actual profit;
• a funnel showing leads through to completed purchases;
• a scatterplot of regional advertising expenditure against sales, with region as the detail field.

The designer avoids a pie chart because there are many product categories and avoids a secondary axis unless the profit-margin line is explicitly labelled. Conditional formatting is applied to target variance, with icons and text in addition to colour.

Outcome. Managers can see the headline position immediately, investigate regions through the matrix, identify monthly changes in the line chart, and examine possible relationships in the scatterplot. The report does not claim that advertising causes sales; it identifies an association requiring further investigation.

Learning points. The case demonstrates that a report is a coordinated analytical system. Each visual has a distinct purpose, and the usefulness of the page comes from the relationship among visuals, measures, filters, and interaction design.

Limitations, Governance, and Contemporary Practice

Visuals can mislead through truncated axes, inconsistent scales, selective date ranges, inappropriate denominators, excessive colour, or ambiguous titles. A dashboard may also encourage users to treat correlation as causation or a target as an objective measure of success.

Governance requires documented measure definitions, certified datasets where appropriate, controlled access, and protection of sensitive fields. Row-level security should be tested with representative user accounts. Personal, financial, health, or location data should be minimised and displayed only when necessary.

Accessibility includes readable contrast, keyboard navigation where supported, descriptive titles, meaningful alt text, consistent tab order, and non-colour indicators. Report authors should test pages on different screen sizes and avoid relying on hover interactions as the only source of information.

Power BI continues to develop through small multiples, on-object interaction, improved accessibility, natural-language experiences, and AI-assisted interpretation. AI-generated summaries can accelerate exploration, but their statements must be checked against the model, filter context, measure definitions, and source data. Human review remains necessary for consequential decisions.

Chapter Summary

Selecting a Power BI visual requires matching visual form to analytical purpose. Tables and matrices support detailed inspection and hierarchy; cards and KPIs communicate headline performance; bar and column charts compare categories; line and area charts show trends; combo charts compare related measures with care; scatterplots reveal relationships; treemaps show hierarchical composition; waterfall charts explain sequential change; and funnels represent ordered stages. Maps and interactive features extend analysis when geography or exploration is genuinely relevant.

Effective configuration depends on valid measures, appropriate aggregation, chronological ordering, meaningful sorting, clear units, accessible colour, and honest scales. Visuals should be tested for interaction behaviour, privacy, performance, and interpretability. A professional Power BI report communicates evidence without overstating what the data can prove.

Key Terms

Visual encoding: The use of position, length, colour, size, or shape to represent data.

Measure: A calculation evaluated within the current filter context.

Dimension: A descriptive field used to group, filter, or categorise data.

Filter context: The set of filters that determines a measure’s result.

Table: A visual showing values in rows and columns.

Matrix: A hierarchical table with row groups, column groups, and subtotals.

Card: A visual displaying a prominent single value.

KPI: An indicator comparing actual performance with a target or benchmark.

Gauge: A progress visual comparing a value with defined limits or a target.

Clustered chart: A chart placing series beside one another for comparison.

Stacked chart: A chart placing series segments within a common bar or column.

Line chart: A visual showing change across an ordered axis.

Combo chart: A visual combining column and line encodings.

Scatterplot: A chart showing the relationship between two numeric variables.

Treemap: A nested-rectangle visual for hierarchical part-to-whole analysis.

Waterfall chart: A visual showing sequential positive and negative contributions.

Funnel: A visual representing quantities across ordered process stages.

Small multiples: Repeated charts using the same scale for different categories.

Drill-through: Navigation to a detail page filtered to a selected context.

Conditional formatting: Formatting that changes according to data values or rules.

Review Questions

Short Answer Questions

1. What is the main analytical purpose of a card visual?
2. Why is a bar chart often preferable when category names are long?
3. What type of question is best addressed by a line chart?
4. State one risk associated with using a secondary axis in a combo chart.
5. What does a funnel chart represent?

Long Answer Questions

1. Compare tables, matrices, cards, and KPIs, including the situations in which each is most appropriate.
2. Explain how sorting, axis selection, data labels, and stacking affect the interpretation of bar and column charts.
3. Discuss the strengths and limitations of line charts and area charts for time-series analysis.
4. Explain how a scatterplot can support segmentation while avoiding unsupported causal claims.
5. Describe a systematic workflow for selecting and configuring a Power BI visual.

Analytical / Application Questions

1. A sales manager wants to compare 30 products, show monthly movement, and monitor performance against budget. Recommend a set of visuals and justify each choice.
2. A report uses a pie chart for 18 categories and a dual-axis chart comparing units and profit margin. Identify the design problems and propose improvements.
3. A funnel shows fewer completed purchases than leads. Explain what additional definitions, measures, and data checks are needed before concluding that the sales process is underperforming.

Multiple Choice Questions with Answers

Question 1

Which visual is most appropriate for displaying exact values for many individual records?

A. Table

B. Gauge

C. Scatterplot

D. Funnel

Correct Answer: A. Table

Question 2

Which visual is primarily designed to show change across an ordered time axis?

A. Treemap

B. Line chart

C. Waterfall chart

D. Matrix

Correct Answer: B. Line chart

Question 3

What is a major risk of using a secondary axis in a combo chart?

A. It prevents all filtering

B. It removes categorical fields

C. It may imply a misleading relationship between differently scaled measures

D. It allows only one measure

Correct Answer: C. It may imply a misleading relationship between differently scaled measures

Question 4

Which visual best explains how several positive and negative effects produce a final total?

A. Waterfall chart

B. Card

C. Map

D. Donut chart

Correct Answer: A. Waterfall chart

Question 5

In a scatterplot, what does each point commonly represent?

A. A page filter

B. An observation or category with values on two numeric axes

C. A report title

D. A subtotal only

Correct Answer: B. An observation or category with values on two numeric axes

Question 6

Which visual is most suitable for an ordered sequence such as leads, proposals, and won deals?

A. Funnel

B. Treemap

C. Gauge

D. Matrix

Correct Answer: A. Funnel

Question 7

Why should month names be sorted using a month-number field in a line chart?

A. To increase the number of measures

B. To prevent alphabetical rather than chronological order

C. To create a secondary axis

D. To enable row-level security

Correct Answer: B. To prevent alphabetical rather than chronological order

Question 8

A 100% stacked chart primarily compares what?

A. Absolute totals only

B. Geographic distance

C. Proportions within totals

D. Individual transaction details

Correct Answer: C. Proportions within totals

Question 9

Which visual is designed for hierarchical part-to-whole analysis using nested rectangles?

A. Treemap

B. KPI

C. Line chart

D. Table

Correct Answer: A. Treemap

Question 10

What is a key accessibility principle for conditional formatting?

A. Use colour as the only signal

B. Avoid all labels

C. Use additional text, icons, or patterns so meaning does not depend only on colour

D. Hide the legend

Correct Answer: C. Use additional text, icons, or patterns so meaning does not depend only on colour