Tableau dashboards, interactions, performance and publishing
Assemble worksheets into interactive dashboards, keep them fast, publish them, and compare the approach with Power BI.
This lesson builds on: Tableau — dimensions, measures, calculations and Level of Detail.
- Lay out a dashboard with containers and sizing
- Add filter, highlight, URL and parameter actions
- Apply Tableau performance practices
- Publish to Tableau Cloud/Server or Tableau Public and manage access
Building a dashboard
New Dashboard → set Size (fixed size, e.g. 1366 × 768, for predictable layout; Automatic resizes unpredictably). Use horizontal and vertical containers to align KPI tiles, charts and filters; tiled layout is easier to maintain than floating objects.
A sound executive layout mirrors the Power BI guidance: KPI row → trend → breakdown → detail, ≤ 6 views, one accent colour.
Dashboard interactions (actions)
Dashboard → Actions:
| Action | Effect |
|---|---|
| Filter | Clicking a mark in one view filters others (e.g. click a store → product chart filters) |
| Highlight | Emphasises related marks without filtering |
| URL | Opens a link (e.g. CRM record) |
| Go to Sheet | Navigation |
| Parameter / Set actions | Clicking updates a parameter or set — powerful for "what-if" and dynamic comparisons |
Add a "Use as filter" funnel icon to a view for quick filter actions. Make interactions discoverable: a short instruction ("Click a store to filter") helps users.
Performance
Tableau's Performance Recorder (Help → Settings and Performance → Start Performance Recording) shows slow queries and rendering.
- Use extracts, filtered to needed rows and columns; aggregate extracts to the visible level where possible.
- Prefer relationships/pre-joined views to complex blends.
- Limit quick filters with many values ("Only relevant values" is expensive); use parameters or action filters instead.
- Avoid huge crosstabs (thousands of marks); use summary views with drill-down.
- Push heavy calculations to the source (SQL views).
Publishing
- Tableau Cloud / Server: publish workbooks and data sources (Server → Publish). Publishing a data source separately lets many workbooks share certified, governed data (like Power BI semantic models). Schedule extract refreshes; on-premises data for Tableau Cloud needs Tableau Bridge.
- Permissions: projects with permission rules for groups; row-level security via user filters or data policies.
- Tableau Public: free, but everything published is public — never publish confidential data there. It is ideal for portfolios using fictional data such as the OrionStats datasets.
Comparative project: Power BI vs Tableau
The same dashboard in both tools (Project 27) teaches what is conceptual and what is tool-specific:
| Aspect | Power BI | Tableau |
|---|---|---|
| Modelling | Explicit star schema, DAX measures | Relationships canvas, calculations, LODs |
| Strength | Business semantic models, Microsoft ecosystem, cost | Visual exploration, flexible visual grammar |
| Calculation thinking | Filter context, CALCULATE | Level of detail, table calculations |
| Sharing | Workspaces, apps, RLS | Projects, permissions, user filters |
Employers value the concepts: grain, star schemas, measure definitions, context, design. OrionStats certifications let you choose either tool for practical work and assess these equivalent competencies.
Practice
True or false: workbooks published to Tableau Public can be viewed by anyone, so it must not be used for confidential company data.
You want clicking a store in a bar chart to filter the product chart beside it. Which dashboard action?
Summary
- Fixed-size, container-based layouts keep dashboards predictable.
- Filter, highlight, parameter and set actions create interactivity.
- Use extracts, limit expensive quick filters, push calculations upstream.
- Publish governed data sources; never put confidential data on Tableau Public.