Selecting and Formatting Appropriate Visuals

This episode focuses on a practical PL-300 objective: Selecting and Formatting Appropriate Visuals. An appropriate visual should make the business question easier to answer, while formatting, labels, units, and emphasis should reinforce rather than compete with the analytical me…

A sales director opens a Power BI report and asks a simple question. Which product category performed best this quarter? The report contains all the necessary numbers, but they appear in a crowded visual with unclear labels, unexplained units, and colors that compete for attention. The visual renders successfully.

Nothing is technically broken. Yet the director still can’t answer the question with confidence. What evidence would show that the analyst chose and formatted the visual correctly, rather than merely creating something Power BI was willing to display?

That distinction is central to this part of the PL-300 blueprint. The stated skills include selecting an appropriate visual and formatting and configuring visuals. Those are applied judgment skills. They aren’t satisfied by remembering where a formatting control appears. You need to understand the question, identify the comparison the consumer must make, choose a visual that preserves that comparison, and then test whether the result can be interpreted without guesswork.

Begin with the required outcome. A request such as create a sales chart isn’t specific enough to drive a good decision. Does the business want to compare categories, follow revenue over time, examine how parts contribute to a whole, inspect a distribution, discover a relationship between two measures, or review detailed records? Each task asks the consumer to notice something different. A visual that works well for one task might conceal the evidence needed for another.

Suppose the director’s actual requirement is to compare quarterly sales across six product categories. The important relationship is category against category. The consumer needs to see which categories are larger, which are smaller, and how substantial the differences are. A straightforward bar or column visual can preserve that comparison.

A table could also contain the correct values, but it asks the director to read and compare individual cells. A single total would be simpler still, but it would destroy the category comparison entirely. Simplicity doesn’t mean removing information that the question requires. It means avoiding complexity that doesn’t help answer the question.

This decision belongs primarily in the report layer. Power Query performs refresh-time data transformations. The semantic model organizes data and relationships, while Data Analysis Expressions, or DAX, can calculate the sales measure.

Those layers determine what data and results are available. They don’t decide how the report consumer should visually compare six categories. The report visual presents the result and responds to the applicable filter context, which is the set of filters affecting the data evaluated for that visual or expression.

Formatting then supports the same analytical task. A title should say what the visual represents. Labels should identify categories or values clearly. The axis and units should make the scale understandable.

If the values represent millions in currency, that meaning shouldn’t be left for the director to infer. Formatting isn’t merely decoration applied after the analytical work. It’s part of making the intended comparison unambiguous.

Now imagine a different request using the same sales data. The director wants to understand whether revenue is rising, falling, or changing irregularly from month to month. That is a trend question rather than a category comparison. A line visual is a natural candidate because it preserves the sequence over time and makes changes between periods easier to follow. Reusing the earlier category visual simply because it already exists would place convenience ahead of the requirement.

This is why there isn’t one universally correct visual. The correct choice depends on the analytical task and the information that must remain visible. Comparison, trend, composition, distribution, relationship, and detail are different forms of inquiry.

A familiar visual can still be the wrong choice if it weakens the relationship the consumer needs to examine. Power BI offering a visual type or formatting option only tells you that the option is available. It doesn’t prove that the option serves this report.

The simplest useful visual is the one that preserves the necessary comparison without adding avoidable interpretation work. For the monthly trend, that means retaining the time sequence and the relevant measure. Adding several unrelated measures might make the report look more comprehensive, but it can also obscure the original question. Removing the time axis would be worse because the consumer could no longer see the progression that defines a trend.

Configuration determines how effectively the selected visual communicates. Consider a chart titled Sales with an axis labeled only one, two, three, and four. Those numbers might be quarters, regions, or internal category codes.

The values might represent dollars, units, or percentages. The underlying calculation could be correct while the presentation remains ambiguous. A specific title, recognizable category labels, an intelligible axis, and explicit units connect the visual to the business meaning.

Troubleshooting should therefore start with an observable symptom. Suppose the monthly revenue line looks correct when no filters are applied, but it becomes confusing when a consumer selects one region. Reproduce that state instead of changing settings at random.

Record what the consumer expected and what appeared. Check the active filter state and the configuration directly governing the visual. Then compare the visual’s values and labels with the filtered result the report is meant to communicate.

The cause might not be the visual type. If the DAX result is wrong under that filter context, replacing the line with bars won’t repair the calculation. If the model doesn’t provide the intended result, changing colors won’t repair the model. Conversely, if the values are correct but the axis, title, or units misrepresent them, rebuilding a Power Query transformation addresses the wrong layer. The symptom must be traced to the layer that controls it.

After making a correction, test the original state and the filtered state again. A successful configuration change isn’t proof by itself. The evidence is the observed behavior. Can the director still read the overall trend?

Does selecting a region produce the expected regional result? Are the title, labels, axis, and units still accurate in that state? Retesting both a normal case and an edge condition helps reveal whether the change fixed the requirement or merely moved the confusion somewhere else.

Consumer interaction adds another dimension to the choice. A report isn’t always viewed in one fixed state. Filters and selections can change the filter context, which can change the values a visual presents. The design must remain interpretable as those values change.

A label that makes sense for a company-wide total may become misleading when the consumer narrows the report to one division. A unit that works for a very large total may make smaller filtered values difficult to distinguish. The relevant question is still whether the configured presentation supports an unambiguous interpretation.

Take a sales-and-operations report with revenue by category, monthly revenue, and order-level detail. These three needs shouldn’t automatically become three copies of the same chart. The category question calls for direct comparison.

The monthly question calls for a visible sequence over time. The detailed question requires individual records and their relevant fields. Treating all three as interchangeable because they use the same dataset overlooks what each consumer is trying to learn.

The same reasoning applies to composition, distribution, and relationship questions. If the requirement concerns contribution to a whole, the visual must preserve the connection between each part and that whole. If it concerns distribution, the consumer needs to see how observations are spread rather than only a total.

If it concerns the relationship between measures, the presentation must allow those measures to be examined together. The task comes first. The visual is chosen because it exposes the evidence needed for that task.

Formatting choices should reinforce that evidence. A title such as Quarterly Revenue by Product Category identifies both the measure and the grouping. Units establish whether a displayed value means dollars, thousands of dollars, or millions of dollars. Axis information gives values a scale.

Labels identify what a mark or position represents. These elements reduce the amount of hidden knowledge required from the consumer. Without them, two people can look at the same visual and reach different interpretations even though the underlying data hasn’t changed.

More formatting isn’t necessarily better. If every category has a different color, every point has a label, and several decorative elements compete with the values, the report may require more effort to read. The decision rule isn’t to use every available option. Configure what the consumer needs for the intended comparison. Formatting earns its place when it clarifies identity, magnitude, sequence, units, or another aspect required by the question.

You can validate that design through consumer behavior rather than personal preference alone. Give the report a precise question such as which category had the greatest quarterly revenue, or in which month did the direction of revenue change. Apply the intended filter state and inspect what the report presents. The relevant value should be present, but presence alone isn’t enough. The consumer should be able to identify what the value measures, what it is grouped by, which units apply, and what filter state shapes the result.

Suppose the category chart passes the unfiltered check. Category A is visibly the largest, and its value agrees with the expected report result. Then a region filter is applied. Category C becomes the largest, but the title still claims to show company-wide sales.

The marks may be numerically correct, yet the presentation now suggests the wrong scope. That is a report interpretation failure. Verification caught it because the analyst examined the experience under the condition in which the report would actually be used.

Now consider the opposite case. The title and units are clear, but the category values don’t agree with the expected filtered result. That points beyond formatting.

Inspect the data result and the filter context before redesigning the visual. A polished chart can faithfully present an incorrect calculation, just as a poorly formatted chart can obscure a correct one. Selection, formatting, and data correctness are connected, but they aren’t substitutes for one another.

This separation helps prevent random troubleshooting. First restate the requirement in concrete terms. Next identify the layer that owns the observed behavior. Then inspect the setting, calculation, relationship, transformation, interaction, or access condition that directly controls it.

For this episode’s main skill, the selection and presentation decisions usually live in the report layer. If the displayed result itself is wrong, the investigation may lead into the semantic model, DAX, or a transformation. Validation happens where the suspected behavior can actually be observed.

It’s tempting to accept a visual when it looks right under the default view. That’s only one condition. Reports can encounter different filters, refreshed data, roles, and interaction states. Not every visual needs an elaborate test program, but important requirements should be checked under conditions likely to change their interpretation.

A category comparison should remain a category comparison when the values change. A trend should retain a meaningful time sequence. Detail should remain identifiable as detail rather than being mistaken for an aggregate.

The nearest alternative is often valid for another question, which is why memorized rules can fail. A table isn’t inherently inferior to a chart. It may be the better choice when the consumer must inspect precise records or several exact values.

A chart isn’t inherently more analytical. It’s useful when position, length, sequence, or another visual relationship makes the required comparison easier. The decision is about preserving the task, not ranking visual types from good to bad.

The same caution applies to reducing everything to a single headline number. A single value can answer a narrow question such as total revenue for the selected period. It can’t, by itself, show which category led, how the total changed over time, or how observations were distributed. If the requirement includes one of those relationships, the headline number has removed necessary evidence. A visually simple result can still be analytically incomplete.

A practical review can follow the consumer’s reasoning. Ask what business question is being answered. Identify whether it concerns comparison, trend, composition, distribution, relationship, or detail.

Confirm that the visual retains the fields and relationships necessary for that task. Check that the title, labels, axes, and units remove ambiguity. Then change a relevant filter or interaction state and confirm that both the result and its presentation still mean what they claim to mean.

This review also exposes decorative fixes that don’t address the cause. If consumers can’t distinguish quarterly values because the units are unclear, changing the background color won’t help. If the visual omits the category needed for comparison, adding a more descriptive title won’t restore that category. If the calculation is incorrect under a filter, changing the visual type won’t correct the calculation. Each correction should correspond to the evidence that failed.

For a governed sales report, that discipline matters because the report’s appearance may be consistent while its interpretation varies across conditions. Governance doesn’t make every visual choice universal. The analyst still has to map each requirement to the appropriate report behavior and verify what consumers receive. A standard style can support consistency, but it can’t decide whether the business question requires a comparison, a trend, or detailed records.

The result you’re looking for is explainable. You should be able to say why the selected visual preserves the needed comparison, why the nearest alternative would make the question harder or remove necessary information, and how the formatting prevents ambiguity. You should also be able to name the evidence used to validate it, such as the expected values under a controlled filter state and the consumer’s ability to interpret the title, scale, labels, and units correctly.

That standard is more reliable than configuration success. Power BI may accept the fields, render the object, and save every formatting choice without warning. Those events prove that a configuration is possible. They don’t prove that it answers the business question. Correctness appears in the connection between the requirement, the visual’s analytical structure, the configured presentation, and the behavior observed when the report is consumed.

Choosing a visual starts with the question, not with a favorite chart or a familiar button. The visual must preserve the type of meaning the consumer needs, such as a comparison, trend, composition, or distribution. Relationships and detailed records call for different visual choices. Its title, labels, axes, and units must make that meaning clear.

Just as importantly, a report that renders successfully hasn’t necessarily been validated. Test the result in the filter and interaction states where people will use it, and investigate the layer that actually controls any failure. When you can explain why the visual fits the requirement and point to verifiable evidence that it remains accurate and understandable, you’ve moved from configuring a report object to making a defensible analytical decision.

Select a visual by matching its encoding to the comparison, trend, distribution, or relationship being communicated. What misreading by a consumer would tell you the visual is wrong even when the numbers are correct? The episode reference preserves the non-narrated details needed for practice and later recall. This lesson was produced by Bare Metal Cyber dot com.

Selecting and Formatting Appropriate Visuals

headphones Listen Anywhere

More Options »
Broadcast by