roadmap
- 01
- 02
- 03
- 04
- 05
- 06
- 07
- 08
Prepare for Tableau interview questions on turning a data source into charts and dashboards. Covers connections, dimensions and measures, LOD calculations, filters, chart choice and publishing.
58 of 58 questions in your plan: 0 known, 0 in progress, 0 due, 0 skipped, 58 new. 0 cut, outside your plan.
A round set aside for Tableau alone is rare. Data analyst and BI analyst loops fold the tool into their own work, usually as a screen-shared build against a dataset you have not seen, or as a take-home dashboard you defend afterwards in a walkthrough.
The build round is graded on what the numbers mean, not on the clicks. A profit ratio written as a row-level division and then averaged disagrees with total profit over total sales, and the candidate has to say which one the business meant. A join from orders to returns quietly doubles the sales of every order with two return lines. A customer's first purchase date calls for a FIXED expression, and a region filter leaves that date untouched until the filter is added to context. Chart questions sit beside these: why a bar has to start at zero when a line need not.
Further up the level range the conversation leaves the worksheet for the server: a live connection or an extract, and what each costs the database behind it; how one published workbook shows each regional manager only their own rows; why a dashboard crowded with quick filters takes seconds to answer a click.
Chart design sits ahead of the tool on purpose: it is tool-neutral, and it supplies the vocabulary the platform module takes for granted. Deciding which field drives which visual channel, whether position, length, color or size, is the decision the Rows and Columns shelves and the Marks card ask for, and the statistic a chart computes for you is what Tableau does the moment it sums a measure you only dragged onto a shelf. The rest of the module, from matching the mark to the comparison to the defaults that mislead, is where the chart questions of a build round come from.
Tableau itself then runs from the data source to the audience. Connections come first, because whether tables are related or joined, and whether the data is live or extracted, fixes what every later sum counts. Worksheets and marks bring dimensions and measures and the level of detail of the view, the idea the next topic is named after: LOD expressions make sense only once you can say what the view already aggregates to.
Filters and parameters come after LOD rather than before it, because the question interviewers return to is sequence: a context filter acts before a FIXED expression is computed, an ordinary dimension filter after it. Dashboards and actions turn those filters into clicks. Publishing closes the plan, where a finished workbook meets permissions, refresh schedules and row-level security on Tableau Server.
An analyst who writes SQL but has never opened Tableau should rebuild one chart per topic in the free Tableau Public as the plan reaches it. Choosing by topics as the study order keeps the questions in the plan's sequence, so the view's level of detail arrives before the expressions that depend on it. Power BI users bring a DAX habit, a measure that rewrites its own filter context, which does not carry over to LOD expressions; for them the calculated fields topic deserves the longest stay. A take-home due in days is the case for leaving level at its full range, since a brief rarely says which seniority it grades, and narrowing importance to must know. The hours saved belong to the dashboard, which the follow-up call takes apart chart by chart. Everything around the tool is outside this plan: the SQL that feeds a data source, the statistics behind an experiment and the business case an analyst loop wraps around a dashboard. The [BI Analyst roadmap](/roadmaps/bi-analyst) is the loop this tool usually sits inside, and the [Power BI roadmap](/roadmaps/power-bi) prepares for the tool a posting may name instead.
An analyst who writes SQL but has never opened Tableau should rebuild one chart per topic in the free Tableau Public as the plan reaches it. Choosing by topics as the study order keeps the questions in the plan's sequence, so the view's level of detail arrives before the expressions that depend on it. Power BI users bring a DAX habit, a measure that rewrites its own filter context, which does not carry over to LOD expressions; for them the calculated fields topic deserves the longest stay.
A take-home due in days is the case for leaving level at its full range, since a brief rarely says which seniority it grades, and narrowing importance to must know. The hours saved belong to the dashboard, which the follow-up call takes apart chart by chart.
Everything around the tool is outside this plan: the SQL that feeds a data source, the statistics behind an experiment and the business case an analyst loop wraps around a dashboard. The BI Analyst roadmap is the loop this tool usually sits inside, and the Power BI roadmap prepares for the tool a posting may name instead.