Tableau / Intermediate: calculation and interaction scope / Lesson 12

Dashboard actions, navigation and accessibility

Design source-to-target interactions whose selected population, clearing behavior and reading order remain explicit.

About 4 min readContent reviewed 2026-10-28

On this page

Lesson goals

Separate filter, highlight and navigation actions.
Define source/target fields and cleared-selection outcomes.
Review non-color, keyboard and device journeys.

Define the reader journey

A dashboard should guide a task such as select a product, compare its regional sales and inspect detail. Name the initial state, selection, target result and return/reset path. Selecting Tea should lead to the intended 42 population, not merely a page that appears to change. Source and target sheets can use different data sources or keys, so map the actual fields and types. A layout is not a data relationship. The first record is a paper interaction contract, and no dashboard, action or navigation is created or clicked in this course.

text Synthetic dashboard action contract

source:Product mark Tea
target:regional Sales for the same Product key
expected:North12 + South30 =42
clear selection:explicitly restore, retain or empty per approved design
return/reset:visible and understandable

The record defines intended states only and captures no Tableau interaction.

Distinguish filter and highlight

A filter action sends selected source values as target filters, while highlighting can emphasize matching marks without removing the rest of the population. The distinction affects totals and interpretation. Specify triggering behavior and source/target field mapping, rather than applying all fields indiscriminately. A multi-mark selection can produce a different allowed set from a single mark. Clearing selection can restore all, keep a filtered state or empty a target according to supported settings. Choose the behavior deliberately and make it understandable; a target that stays filtered after the source looks clear can otherwise mislead.

Make context and navigation visible

A detail destination should display the selected entity, date scope and additional filters. A navigation action does not automatically carry every expected filter or prove the source and target represent the same business key. Test direct entry without source context as well as normal navigation. Keep a clear back/reset path. A hidden sheet or obscure destination is not a security control; consumer permissions and row restrictions require a separate governed design. Avoid assuming that a dashboard’s current state describes all other workbook sheets or export/download capabilities available to the same user.

Model clear-selection outcomes

The second program compares Tea filtering with an all-products reset using synthetic facts. It establishes 42 versus 60 only. It does not implement highlighting, action scope or state persistence. Review empty and multi-selection behavior separately. Labels, captions and meaningful descriptions should communicate state without color alone. Keyboard order and supported assistive access need actual review; decorative objects can distract from the task or create confusing focus. Mobile/device layouts can change available space and interaction affordances, so the same analytical contract should be checked through each supported reader route rather than assumed from a large authoring screen.

python Synthetic filter and reset oracle

rows = [("Tea", 12), ("Coffee", 18), ("Tea", 30)]
selected = {"Tea"}
print(sum(amount for product, amount in rows if product in selected))
print(sum(amount for _, amount in rows))

Expected output

42
60

42 models Tea and 60 the full fixture. No action, UI or accessibility test is performed.

Verify the complete experience

Review default, one selection, multiple selections, clear, direct detail entry and no-data cases. Check expected facts and amounts, visible context, return behavior and non-color interpretation. This course creates no dashboard, tests no browser and publishes no content. Its subset calculation is a local oracle only. A future authorized review should record mouse, keyboard and device observations and any accessibility limitations. A complete interaction handoff describes the source, target, matching fields, trigger, state transition and reader evidence, so an attractive button cannot stand in for correct analytical behavior.

Common errors and fixes

Highlighting is interpreted as removal from the total

Why it happens: Visual emphasis was confused with filtering.
How to fix it: Inspect action type and explain selected versus full population.
Prevent it next time: Test a known selected amount and overall amount together.

Clearing a source leaves an unexplained target filter

Why it happens: Clear-selection behavior was not specified.
How to fix it: Choose and label the approved restore/retain/empty outcome.
Prevent it next time: Keep clear and multi-select cases in acceptance.

An attractive desktop layout is assumed accessible on every device

Why it happens: Supported consumption routes were not observed.
How to fix it: Review keyboard, non-color and device-specific journeys separately.
Prevent it next time: Record actual accessibility observations and remaining limits.

Check your understanding

1. Why should clearing Tea be a separate test from selecting it?

Show worked solution

Selecting can correctly filter 42 while clearing unexpectedly retains 42 or empties the target. The intended reset might be 60. Define the state transition and visible explanation, then validate it with known facts. A source without a selected mark does not automatically prove that every target has returned to the initial state.

2. What must a detail page explain after product navigation?