Creating Custom Tooltips and Managing Visual Layers
A sales report contains a chart showing revenue by product category. The business wants a reader to get extra context about a category without filling the report page with more charts. The same page also contains a title, a decorative background, several data visuals, and a transparent object that supports an interactive design. Everything appears correct until the analyst tries to select one visual and keeps grabbing another object instead.
Both requirements belong to the report layer, but they solve different problems. A custom tooltip provides contextual detail without permanently taking up space on the main canvas. The Selection pane helps the report author organize objects and control their grouping, visibility, and front-to-back order, which is also called z-order. One feature changes what contextual information the consumer receives. The other helps determine how report objects are organized and layered.
These are part of the current Power BI Data Analyst skill set. The published scope includes creating custom tooltips and grouping and layering visuals with the Selection pane. That wording points toward applied report design. It isn’t enough to know that the features exist. You need to connect each one to a requirement, understand where its behavior is controlled, and test whether the consumer gets the intended result.
Suppose the sales team asks for margin, units sold, and the most recent reporting period to appear as supporting context for each product category. Those values matter, but the team doesn’t want three additional visuals competing with the main chart. A tooltip is a reasonable candidate because the information is supplementary. The report can preserve its main layout while making the extra context available through the tooltip interaction.
Now compare that with a different request. The sales director says that the category name and revenue must always be visible because those values drive the primary decision. Hiding them inside a tooltip would be a poor design choice.
A user might never invoke the tooltip, and accessibility guidance warns against making a tooltip the only carrier of essential information. The appropriate question isn’t whether Power BI permits a value to appear in a tooltip. It’s whether that value should be optional context or part of the report’s persistent message.
That distinction gives you a useful decision rule. Put essential information where the consumer can encounter it as part of the normal report experience. Use a custom tooltip for helpful detail that adds meaning without becoming the only way to understand the result. A configuration can be technically functional and still fail that usability requirement.
The same requirement-first thinking applies to the Selection pane. Imagine that the analyst wants to place a text label over a colored shape while keeping a chart in front of the page background. This isn’t a data-transformation problem. Changing Power Query won’t decide which object sits in front.
A relationship in the semantic model won’t make a background shape easier to select. A Data Analysis Expressions calculation, usually called DAX, can calculate a value, but it doesn’t control the visual stacking of report objects. Grouping, visibility, and z-order are report-layer concerns, so the Selection pane is the relevant feature.
Separating those layers prevents wasted troubleshooting. Power Query is where refresh-time transformations shape query output. The semantic model organizes data and relationships.
DAX evaluates calculations under the applicable filter context, meaning the set of filters affecting an expression or visual. Custom tooltips and Selection-pane organization operate in the authored report experience. Publishing a report to the Power BI service doesn’t turn a report-layer design fault into a query or model fault.
A custom tooltip needs more than a collection of interesting numbers. Its information must make sense for the context in which the consumer encounters it. Return to the category chart. If the consumer invokes the tooltip for the Bikes category, the expected supporting information should describe Bikes under the relevant report state. If the tooltip presents values that appear unrelated to that category or to the active filters, something about the result is wrong, even if the tooltip itself opens successfully.
This is where filter context becomes practically important. The visible result is evaluated under filters affecting that interaction. A report might already be limited to a region, a quarter, or a sales channel.
The tooltip’s values need to be interpreted in that context. You don’t prove correctness by seeing any number appear. You prove it by predicting what the number should represent, applying a controlled report state, and comparing the observed result with that prediction.
For example, suppose the report is filtered to the West region and the consumer examines the Bikes category. The expected tooltip should provide the approved supplementary measures for Bikes within the applicable state of the report. If the values instead appear to represent all regions, the analyst should reproduce that condition and investigate the configuration governing the result. Randomly changing Power Query steps or rearranging unrelated visuals would avoid the actual question.
The fault could concern how the report interaction supplies context, how the underlying model responds to that context, or how a calculation evaluates it. Those possibilities cross connected Power BI layers, but the symptom still needs to be described precisely. Does the tooltip fail to appear at all?
Does it appear but show an unexpected value? Is the value correct without a regional filter and incorrect with one? Each observation narrows the investigation.
A normal case alone isn’t enough. If the unfiltered report produces a plausible total, that doesn’t prove that filtered conditions work. Retest with at least one meaningfully different state.
Choose a category with known expected behavior, apply a controlled filter, invoke the tooltip, and compare the displayed details with the visible report context. Then remove or change the filter and confirm that the result responds as expected. The evidence is the consumer-visible behavior, not the fact that an authoring control accepted the configuration.
The Selection pane needs the same disciplined validation. Consider a page containing a background shape, a chart, a label, and another object intended to sit above part of the chart. If an object that should be behind the chart is actually in front, it may intercept selection or obscure content. The canvas can look almost correct while the interaction remains unreliable. The relevant evidence is which object is visible, which object can be selected, and whether their front-to-back order matches the intended design.
The Selection pane exposes report objects in a form that lets the author reason about those layers. Z-order determines which overlapping object is in front of another. Visibility determines whether an object is currently shown.
Grouping lets related objects be managed together. These controls don’t alter the sales data or rewrite a measure. They alter the composition and manageability of the report page.
Naming also matters once the page becomes complex. A page with many unnamed or indistinguishably named objects forces the author to identify them by trial and error. Meaningful names make the relationship between the pane and the report design easier to understand. A group can represent objects that belong together, such as a label and its background, rather than leaving a future maintainer to infer that relationship from their position alone.
Grouping doesn’t mean every nearby object should be treated as one unit. The grouping should reflect how the objects need to be maintained. If a decorative shape, title, and supporting label are designed to move or be managed together, grouping may express that intention. If a data visual needs separate control, putting it into the same group merely because it overlaps could make later changes more awkward. The interface allows multiple arrangements because report requirements differ.
Picture the sales report with two visual states built from overlapping objects. An analyst can make the page look correct in one state but leave an unwanted object visible in another. A quick glance at the initial state won’t reveal the defect. Validation has to exercise the relevant interaction state and observe which objects remain visible and which layer appears in front. If visibility or grouping participates in a larger interactive design, test every state that the business expects consumers to use.
This is why tidy naming and grouping aren’t just cosmetic habits. They reduce ambiguity during troubleshooting. If an unexpected panel covers the category chart, a clearly named object or group gives the analyst a direct place to investigate. With a disorderly list of similar objects, the analyst may hide or reorder the wrong item, create a temporary improvement, and introduce another failure elsewhere on the page.
The consumer’s experience remains the deciding evidence. An analyst might look at the Selection pane and conclude that the order appears reasonable. That’s only a configuration inspection.
The analyst should also use the report as a consumer would. Check that intended content remains visible, decorative elements don’t cover data, and overlapping objects behave correctly under the expected interactions. Configuration and observed behavior should agree.
Now combine the two features on the same sales page. The category chart uses a custom tooltip to provide supplementary margin and unit information. A shaped background and label frame part of the design. The analyst has grouped the frame elements and positioned that group behind the chart. In the standard report state, the layout looks clean, and the tooltip appears to work for the overall data.
Apply a regional filter and two faults become visible. The tooltip still displays a value, but it doesn’t match the expected regional interpretation. At the same time, an object that should remain behind the chart is placed above it in one interaction state. The temptation is to call the whole report broken and start changing several settings. A better approach is to split the observations according to the mechanisms that own them.
For the tooltip, record the expected and actual contextual result. Keep the report state controlled. Check the report interaction and the connected model or DAX behavior that determines the value under the active filter context. Correct the direct cause, then repeat the original unfiltered case and the filtered case. The tooltip isn’t validated until both produce the intended contextual meaning.
For the overlapping object, inspect the relevant object or group in the Selection pane. Check visibility, membership in the intended group, and z-order. Make the correction there rather than changing a data transformation or calculation. Then exercise the page state that originally exposed the problem and confirm that the right content remains in front. Return to the normal state as well, because a fix for one state shouldn’t silently damage another.
Notice what this test does and doesn’t establish. Seeing the tooltip appear establishes that the interaction can produce a tooltip. It doesn’t establish that the information is correct, useful, or accessible as supplementary content.
Seeing the page look correct in one state establishes that one arrangement appears acceptable. It doesn’t establish that visibility and layering remain correct in every required state. Each claim needs evidence at the layer where the behavior occurs.
Accessibility adds another reason to be careful with tooltip content. A custom tooltip can enrich the report, but it shouldn’t be the sole route to information the consumer must have. If the sales team must know that a result is provisional, for example, hiding that status only in a tooltip would make the core message dependent on discovering and using that interaction. The design should keep essential meaning available through the normal report experience and reserve the tooltip for added context.
This doesn’t make tooltips unimportant. It gives them a clearer job. They can reduce permanent canvas demand while supporting deeper interpretation.
The discipline is to decide what belongs in the main view and what can remain optional. A crowded canvas isn’t automatically more informative, and a sparse canvas isn’t automatically more usable. The requirement and the consumer’s ability to understand the report determine the balance.
There also isn’t one universal Selection-pane arrangement that is correct for every report. A simple page may need very little grouping. A page with many overlapping and interactive objects benefits more from deliberate names and groups. The goal isn’t to maximize the number of groups or to create an elaborate hierarchy for its own sake. The goal is to make the page’s composition understandable and maintainable while preserving the intended consumer behavior.
When a report passes its first visual check, ask what condition could expose a hidden defect. For a tooltip, use a different filter or category and confirm that the context changes correctly. For layering, test an interaction state where visibility or overlap matters. If roles, refresh conditions, or consumer access can affect the broader scenario, reproduce the specific condition associated with the reported failure rather than assuming the author’s normal view represents every user experience.
Keep the diagnosis proportional to the symptom. An incorrect tooltip value may require investigation beyond the report formatting because its value can depend on model behavior and DAX evaluation. A tooltip that carries essential information is primarily a design and accessibility problem even if its numbers are accurate.
An obscured chart points toward layering or visibility. A confusing list of objects points toward naming and grouping. Similar-looking failures can have different owners.
The current certification scope rewards that distinction. Creating a custom tooltip means more than finding a control. Grouping and layering visuals means more than moving an item once. In each case, you’re translating a requirement into report behavior, preserving the surrounding architecture, and demonstrating that the consumer receives the intended result.
For the sales report, the final acceptance check can be stated in observable terms. The main chart carries the information required for ordinary interpretation. Its custom tooltip adds approved supporting context without becoming the only source of essential meaning.
Under controlled filters, the tooltip’s values reflect the expected context. Named and logically grouped objects remain maintainable in the Selection pane. Their visibility and z-order produce the intended page in each required interaction state.
That standard is stronger than saying the settings were saved successfully. It links the business requirement to the Power BI layer, the layer to its mechanism, and the mechanism to evidence. When those links are clear, troubleshooting becomes less about trying familiar controls and more about testing a specific explanation for what the consumer sees.
Custom tooltips and the Selection pane both shape the report experience, but they answer different needs. A tooltip supplies supplementary context without permanently occupying the main canvas. The Selection pane makes overlapping objects manageable through deliberate naming, grouping, visibility, and z-order.
Neither feature is correct merely because its configuration can be saved. Keep essential meaning out of tooltip-only interactions, organize page objects according to how they must be maintained, and test the result under the conditions consumers will actually encounter. When you can connect the requirement to the owning layer and then prove the expected behavior, you’re evaluating the report as an analyst rather than treating authoring controls as evidence of success.
Explain when a custom tooltip adds context and when the Selection pane clarifies layering. What hover, focus, or overlap condition would you test with a consumer? The matching episode reference turns the spoken reasoning into a compact set of terms and hands-on tests. This lesson was produced by Bare Metal Cyber dot com.