Skip to content
Tutorials Center
My learning

Topics

Lesson 34 of 35 ยท Practice

Tableau Interview Questions and Answers

Prepare for a Tableau interview: concepts, data connections, calculations, LOD, filters, dashboards, performance and security, with short answers and runnable checks.

Tableau interviews usually mix three things: concepts ("what is the difference between a live connection and an extract?"), scenarios ("a dashboard is slow, what do you do?") and calculations ("write a formula for..."). This lesson gives short, accurate answers to the questions that come up most, grouped by topic. Use them to refresh your memory, but practise saying each one in your own words: interviewers can tell the difference between understanding and recitation.

Tableau does not run in the browser, so formulas are read-only. Two of the trickiest topics, level of detail and set vs group logic, include runnable stand-ins in SQL and pandas that show the behaviour on tiny data.

Basics

What is Tableau, and what makes it different from a spreadsheet? Tableau is a visual analytics tool: you drag fields onto shelves and it generates queries and charts. Unlike a spreadsheet it separates data from the view, connects to many sources, handles large data through extracts and supports interactive dashboards.

What are the main products? Tableau Desktop (authoring), Tableau Prep (data preparation), Tableau Cloud and Tableau Server (sharing and governance), Tableau Public (free, public-only) and Tableau Mobile. See Publishing to Tableau Cloud and Server.

What are dimensions and measures? Dimensions are categorical fields that split the data (region, date); measures are numeric fields that are aggregated (sales). Tableau assigns them by data type, and you can convert a field. Blue pills are discrete and green pills are continuous: that is a separate choice (discrete = headers, continuous = axis).

What are the shelves and the Marks card? Columns and Rows define the structure; Filters, Pages, and the Marks card (Color, Size, Label, Detail, Tooltip, Shape) control encodings. Detail adds granularity without a visual encoding.

Workbook vs worksheet vs dashboard vs story? A workbook is the file (.twb, or .twbx packaged with data). A worksheet is one view. A dashboard combines several. A story is a sequence of views (Stories and Word Clouds).

Data connections

Live vs extract? Live queries the source each time (fresh data, depends on source speed); an extract is a compressed .hyper snapshot, fast and offline-capable, refreshed on a schedule (Extracts and Performance Tuning).

Join vs blend vs relationship vs union? A join combines tables row-wise at the data-source level (inner, left, right, full) and can duplicate rows. A union stacks tables with the same columns. A relationship (2020.2+) keeps tables separate and joins only the fields a view needs, preserving each table's level of detail. A blend combines two separate data sources on a linking field after aggregating the secondary.

Why might a join inflate totals? One-to-many fan-out; see Common Tableau Mistakes.

What is a published data source and why use one? A central, governed source with shared logic and permissions, refreshed once.

Calculations

Types of calculations? Basic (row-level and aggregate), table calculations, level of detail (LOD) expressions. Aggregates run in the data source; table calculations run in Tableau on the marks in the view.

What are LOD expressions? They compute an aggregation at a level of detail different from the view. FIXED computes at exactly the dimensions named, ignoring the view; INCLUDE adds dimensions to the view's level; EXCLUDE removes dimensions. FIXED is computed before dimension filters, INCLUDE/EXCLUDE after.

Python FIXED, INCLUDE and EXCLUDE in pandas

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Table calculation addressing vs partitioning? Addressing fields (the Compute Using direction) define how the calculation moves along the data; partitioning fields define the groups in which it restarts.

RANK vs RANK_DENSE vs RANK_UNIQUE? Ties share a rank and the next is skipped (RANK), not skipped (DENSE), or each gets a unique rank (UNIQUE).

What does ATTR do? It returns the value if the field has one value in the view, otherwise *. It is useful for putting a dimension on a text mark in an aggregated view.

Write a calculation for year-to-date sales. One option: SUM(IF [Order Date] <= TODAY() AND YEAR([Order Date]) = YEAR(TODAY()) THEN [Sales] END). Alternatives use DATETRUNC and DATEDIFF. See Calculation Practice with Solutions.

Filters

Types of filters in order? Extract, data source, context, set / conditional / Top N, FIXED LODs, dimension, INCLUDE/EXCLUDE LODs, measure, table calculation. Details: Order of Operations and Filter Pitfalls.

What is a context filter and when to use it? A filter promoted to run first, creating a temporary subset. Use it so that Top N, sets or FIXED calculations respect it. It can slow things, so use sparingly.

Why does my Top N ignore my region filter? Top N is evaluated before ordinary dimension filters; make the region filter a context filter.

Sets, groups, hierarchies, parameters

Set vs group? A group combines members of one dimension into a new category (static unless edited). A set is a custom subset of the data defined by members or a condition, and returns IN/OUT, so it can be dynamic (Top 10 customers) and used in calculations and combined sets. A set can be based on several fields.

SQL A dynamic set as an IN/OUT flag

Press Run (or Ctrl + Enter) to see the result here.

What is a parameter? A workbook variable (number, string, date, boolean) that a calculation, filter, reference line or set can use, controlled by the user (for example, a "Top N" slider or a measure selector). It is not tied to a field's values and does not update with the data unless you use a dynamic parameter.

What is a hierarchy? An ordered drill path such as Category > Sub-Category > Product.

What is a bin? An equal-width bucket on a measure that gives a dimension, as in histograms.

Charts and design

When do you use a dual axis? For two measures with a related meaning, such as sales (bars) and margin (line). Synchronise the axes if the units match; avoid when the pairing is arbitrary.

What is a bullet graph / waterfall / Gantt / treemap? A bullet graph compares an actual to a target with bands, a waterfall shows cumulative contributions (built with Gantt bars), a Gantt shows durations (start and size), a treemap shows part-to-whole with nested rectangles.

How do you pick a chart? By the question: comparison (bar), trend (line), distribution (histogram, box plot), relationship (scatter), part-to-whole (stacked bar, treemap), geography (map). Show Me suggests options but does not know your message.

How do you make a dashboard interactive? Filter, highlight and parameter actions, quick filters, URL and sheet-swapping via containers, and tooltips with viz-in-tooltip.

Dashboard actions? Filter (selection on one sheet filters others), Highlight, URL, Go to Sheet, Parameter and Set actions.

Performance, security, publishing

A dashboard is slow. What do you do? Run the Performance Recorder, find whether time goes to queries, layout or calculations, then use an extract, reduce filters and sheets, fix a high-mark view, simplify calculations and set a fixed size (Extracts and Performance Tuning).

How do you implement row-level security? A data source filter built on USERNAME() or ISMEMBEROF(), an entitlement table or data policies, applied in the published data source (Permissions and Row-Level Security).

Tableau Cloud vs Server? Cloud is hosted and upgraded by the vendor and needs Bridge for private data; Server runs on your own infrastructure with full control.

What is Tableau Bridge? A client installed in your network that lets Cloud refresh extracts and run live queries against private data.

How do you embed a dashboard securely? The Embedding API with Connected Apps, using a signed JWT per user (Exporting, Subscriptions and Embedding).

What is a Tableau packaged workbook? A .twbx zip containing the workbook, extracts and local files, so it opens elsewhere.

Scenario questions

Profit is flat while sales grow: how do you investigate? Break down by discount, product and region; waterfall from list sales to profit; scatter of discount vs profit (Project: Profitability Analysis and Forecast Story).

The total on your view does not match the database. Check join fan-out, filters (data source, context), data blending, null handling, time zones and whether the extract is stale. Reconcile with an independent query.

A user sees data they should not. Check where the filter lives (data source vs worksheet), permissions rules, group membership with Effective Permissions and whether the embedded credentials are shared.

How do you handle a requirement that keeps changing? Prototype in Desktop, validate with the user's real questions, parameterise what varies and keep logic in the published data source.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. Which statement about FIXED LOD expressions is true?

    FIXED expressions ignore the view's level of detail and are evaluated before dimension filters (but after context filters).
  2. 2. What is the main difference between a set and a group?

    A set returns IN or OUT for each row and can be conditional or Top N; a group relabels members of a dimension.
  3. 3. A viewer on Tableau Cloud needs data from an on-premises database to refresh. What do you install?

    Bridge connects Tableau Cloud to private-network data.
  4. 4. A Top 10 filter and a region filter give wrong results together. First fix?

    Top N runs before ordinary dimension filters; a context filter forces the region filter earlier.
  5. 5. What is the best answer to 'the dashboard is slow'?

    Measure first: queries, layout and calculations each have different fixes.

Summary

  • Know the vocabulary: dimensions/measures, discrete/continuous, live/extract, join/blend/relationship.
  • Be able to explain FIXED, INCLUDE and EXCLUDE, table-calculation addressing and the order of operations.
  • Describe a clear process for slow dashboards, wrong totals and security problems.
  • Always back an answer with an example from a real or practice project.

Next: Tableau Cheat Sheet