Tableau Interview Questions and Answers
Last updated 2026-11-19
Tableau interviews ask about the data connection, calculated fields, level of detail (LOD) expressions, the order of operations and dashboard performance. The questions that separate strong candidates are the ones about why a filter does not seem to work, or why an average looks wrong, because both come from not knowing how Tableau evaluates a view.
Each answer below gives the idea and the formula where one applies. For the equivalent modelling skills in Microsoft's tool, see Power BI interview questions; for the wider data analyst process read how to crack the data analyst interview.
| Typical rounds | Calculations and LOD questions, dashboard walkthrough, SQL, managerial |
|---|---|
| Core skills | LOD expressions, table calculations, data modelling, dashboard design |
| Questions here | 14 worked questions |
Ideas that unlock most questions
- Dimensions slice the data and measures are aggregated; pills are discrete (blue) or continuous (green).
- A view is a query at the level of detail set by the dimensions on the shelves.
- Filters, sets, LOD expressions and table calculations are applied in a fixed order.
- Extracts are fast snapshots; live connections show current data but depend on the source.
- Aggregating a ratio is different from averaging row-level ratios.
The order of operations, in sequence
Extract filters, data source filters, context filters, FIXED LOD expressions, dimension filters, INCLUDE and EXCLUDE LOD expressions, measure filters, then table calculations. Remembering this list answers the common question of why a Top N filter ignores another filter: ordinary dimension filters run after context filters and FIXED LODs, so promote a filter to context if it must apply first.
Tableau Interview Questions and Answers: questions with worked answers
1 Discrete versus continuous fields: what is the difference?
easyDiscrete fields (blue pills) produce headers and separate panes, with one mark per distinct value. Continuous fields (green pills) produce an axis with a range. A date can be either: a discrete month gives a header for each month name, while a continuous month gives a time axis that connects points in order. Whether a field is a dimension or a measure (what it is) is separate from whether it is discrete or continuous (how it is shown).
2 Live connection versus extract: when do you use each?
easyA live connection queries the source each time, so data is current but performance depends on the database and its load. An extract is a compressed columnar snapshot (a Hyper file) stored with the workbook or on the server, refreshed on a schedule. Extracts are usually much faster, work offline and reduce database load, but data is only as fresh as the last refresh. Use live for real-time needs and well-tuned databases, and extracts otherwise, with incremental refresh for big tables.
3 Compare joins, relationships and blending.
mediumA join combines tables at the row level before analysis, and can duplicate rows when keys repeat. Relationships (the logical layer) keep tables separate and join them at query time, only for the fields in the view, at the right level of detail, so measures are not duplicated. Blending combines data from separate data sources, aggregating the secondary source at the linking fields and then attaching it to the primary, which behaves like a left join. Prefer relationships, use joins for specific row-level needs and blends when sources cannot be combined.
4 What is the difference between FIXED, INCLUDE and EXCLUDE LOD expressions?
hardFIXED computes at exactly the dimensions you list, independent of the view, and is evaluated before dimension filters. INCLUDE adds dimensions to the view's level of detail, while EXCLUDE removes dimensions from it, and both respond to dimension filters. Use FIXED for stable values, such as each customer's first order date. Use INCLUDE when you need a finer-grained calculation inside a coarser view, such as the average of each customer's total sales shown by region.
text{ FIXED [Customer ID] : MIN([Order Date]) } // first order per customer { INCLUDE [Customer ID] : SUM([Sales]) } // average this in a region-level view { EXCLUDE [Sub-Category] : SUM([Sales]) } // sales ignoring sub-category split5 How do you calculate the percent of total of sales by region?
easyEither use a quick table calculation (percent of total) on the sales measure, or write the calculation yourself. In table-calculation form, SUM([Sales]) / TOTAL(SUM([Sales])) divides each region by the total across the table, and the Compute Using setting decides the direction. A LOD alternative, SUM([Sales]) / SUM({ FIXED : SUM([Sales]) }), does not depend on the layout, and so survives a change of view.
textSUM([Sales]) / TOTAL(SUM([Sales])) // table calculation SUM([Sales]) / SUM({ FIXED : SUM([Sales]) }) // LOD version6 Table calculation versus LOD expression: how do you choose?
mediumTable calculations run on the aggregated results already in the view, after the query returns, so they depend on the layout and the Compute Using direction. They suit running totals, ranks, moving averages and percent of total. LOD expressions run in the data query at a specified granularity and can be used as dimensions or filters, so they suit cohort definitions and per-customer values. If the result must not change when you rearrange the view, use LOD.
7 Create a running total and a 3-month moving average.
mediumBoth are table calculations. Right-click the measure and add a quick table calculation for running total, or write RUNNING_SUM(SUM([Sales])). A moving average of the current and two previous months is WINDOW_AVG(SUM([Sales]), -2, 0). The first two months average fewer points, so note this on the chart. Set Compute Using to the date field so it moves along the time axis, not down a different dimension.
textRUNNING_SUM(SUM([Sales])) WINDOW_AVG(SUM([Sales]), -2, 0)8 Profit ratio: why is SUM([Profit]) / SUM([Sales]) different from AVG([Profit] / [Sales])?
mediumThe first divides the totals, giving a weighted ratio; the second averages each row's ratio, giving every row the same weight regardless of size. With two orders, profit 10 on sales 100 and profit 5 on sales 10, the ratio of sums is 15 / 110 = 13.6 percent, but the average of the row ratios is (10 percent + 50 percent) / 2 = 30 percent, because the tiny order gets equal weight. Business ratios are almost always ratios of sums.
9 Parameters, filters and sets: what is the difference?
mediumA filter removes data from the view. A parameter is a user-controlled single value that does nothing until a calculation, filter or reference line uses it, for instance to select the metric shown or to set a Top N. A set is a custom subset of members, either fixed (chosen by hand) or dynamic (defined by a condition, so membership updates with the data), and returns In or Out as a Boolean. Parameters are not tied to the data source and do not update automatically when data changes.
10 How would you build a cohort retention view?
hardDefine each customer's cohort month with a FIXED LOD on the first order date, compute the months since the cohort using DATEDIFF, and count distinct customers per cohort and age. Divide by the cohort size, the count at month zero, to get retention. The calculations are below, with the cohort on rows, months since first order on columns and the retention percent as colour.
textCohort Month = { FIXED [Customer ID] : DATETRUNC('month', MIN([Order Date])) } Months Since = DATEDIFF('month', [Cohort Month], DATETRUNC('month', [Order Date])) Active = COUNTD([Customer ID]) Cohort Size = { FIXED [Cohort Month] : COUNTD([Customer ID]) } Retention % = COUNTD([Customer ID]) / MAX([Cohort Size])11 How do you optimise a slow dashboard?
mediumReduce the work done per view: use extracts, hide unused fields, aggregate data at the right level and filter at the source with extract and data source filters. Limit the number of sheets and marks per dashboard, avoid quick filters with very high cardinality and use "Apply" buttons for multi-select filters. Prefer Boolean or numeric calculations over string ones, use fixed dashboard sizes so layouts are cached, and use the Performance Recorder to find the slow queries and layouts.
12 Describe dashboard actions and a simple security approach.
mediumFilter actions let selecting marks in one sheet filter others; highlight actions emphasise related marks; URL actions open an external page; parameter and set actions change values from a click. For row-level security use a user filter or a calculated field with USERNAME() or ISMEMBEROF() that compares the signed-in user to an entitlement table, applied as a data source filter so users cannot remove it. Test with the "View as user" feature on Server or Cloud.
text[Owner Email] = USERNAME()13 Dual axis versus combined axis, and when would you avoid a dual axis?
easyA combined (blended) axis puts several measures on one shared axis, which only makes sense when they share units. A dual axis places two measures on separate left and right axes, letting you mix a bar and a line, such as sales and margin percent. Avoid it when the two scales invite false visual correlation, when axes are not synchronised for the same unit, or when a simple pair of aligned charts would be clearer.
Frequently asked questions
Is Tableau Desktop or Prep tested?
Interviews focus on Desktop: calculations, LODs, dashboards. Prep may come up briefly for cleaning, unions and joins.
How do I prepare a Tableau portfolio?
Publish two or three dashboards to Tableau Public with a clear question, a tidy layout and a short write-up of the decision they support.
Do I need SQL for Tableau roles?
Usually yes, for custom SQL, joins and checking your data. See the SQL interview guide for practice.
Tableau or Power BI?
Choose by the job description. The data modelling and chart design habits you learn transfer well between the two.
Related guides
More in Role-wise Interview Questions by Job and Tech Stack
- Terraform Interview Questions and Answers
- Web Performance Interview Questions and Answers
- Web Security and OWASP Interview Questions and Answers
- A/B Testing Interview Questions and Answers
- Android Developer Interview Questions and Answers
- Angular Interview Questions and Answers (Angular 18 and Later)
- Ansible Interview Questions and Answers
- API Testing and Postman Interview Questions
- Automation Testing Interview Questions and Answers
- AWS Interview Questions and Answers
- Azure Interview Questions and Answers
- Backend Developer Interview Questions and Answers