Tableau is a leading data visualisation tool used by analysts and BI developers. These Tableau interview questions cover dimensions and measures, chart types and filters for freshers, then calculated fields, LOD expressions, table calculations, data blending, performance and Tableau Server for experienced developers.
30 questions · Updated
Fundamentals, definitions and simple scenarios. Good for freshers and warm-ups.
Tableau Desktop is used to build visualisations and dashboards.
Tableau Prep cleans and shapes data.
Tableau Server (self-hosted) and Tableau Cloud (hosted) are for publishing, sharing, permissions and scheduled refreshes.
Tableau Public is free but saves work publicly, and Tableau Reader views packaged workbooks.
Dimensions are qualitative fields such as region, category or customer name, used to group and slice data.
Measures are quantitative fields such as sales or profit, which are aggregated.
Separately, discrete fields (blue) create headers and distinct values, while continuous fields (green) create axes with a range.
A date can be either.
A live connection queries the data source directly every time the view changes, so data is always current but speed depends on the database. An extract (.hyper) is a compressed snapshot stored by Tableau, which is usually much faster and works offline, but it must be refreshed on a schedule to stay current.
Bar charts compare categories, line charts show trends over time, maps show geographic patterns, scatter plots show relationships between measures, histograms show distributions, highlight tables and heat maps show patterns in a grid, treemaps show parts of a whole with many categories, and bullet charts compare actuals with targets. Show Me suggests charts based on the selected fields.
The Marks card controls how data is drawn: the mark type (bar, line, circle, square, area, map and so on) and which fields are encoded as colour, size, label, detail, tooltip, shape and path. Dragging a field to Detail adds granularity without changing appearance, and multiple Marks cards appear when using several measures on a dual axis.
A worksheet contains a single view or chart.
A dashboard combines several worksheets, filters, legends, text and images on one screen, with interactivity between them.
A story is a sequence of sheets or dashboards arranged as story points with captions, used to guide an audience through an analysis step by step, like a presentation.
A parameter is a user-controlled value, such as a number, date or string, that can be used in calculations, filters, reference lines and Top N sets.
For example, a parameter can let users choose which measure a chart displays, set a target value or select the number of top customers to show.
Parameter actions let users set values by clicking on marks.
Groups combine dimension members into higher-level categories, such as grouping states into regions.
Sets define a subset of data based on conditions or selection, such as the top 10 customers by sales, and can compare in and out members.
Hierarchies organise related fields, such as category, sub-category and product, so users can drill down with plus and minus icons.
Sorting can be done by clicking axis or header sort icons, through the field's sort options (by data source order, alphabetically, by a field or manually), and can be nested. Filters are added by dragging fields to the Filters shelf, and quick filters (filter cards) let users filter interactively with lists, dropdowns, sliders or relative dates.
A dual axis places two measures on separate axes overlaid on the same chart, each with its own mark type and scale, such as sales as bars and profit ratio as a line; the axes can be synchronised. A blended axis places several measures on one shared axis using Measure Values, which suits measures with the same units and scale.
Tableau Public is a free version where workbooks are saved to a public profile on the web, which is excellent for building a portfolio but must never contain confidential data. Organisations share work securely by publishing to Tableau Server or Tableau Cloud with permissions, subscriptions and embedded views, or with packaged workbooks for Tableau Reader.
Applied problems, trade-offs and questions about your own projects.
Extract filters, then data source filters, then context filters, then FIXED level of detail calculations, then dimension filters, INCLUDE and EXCLUDE LODs, measure filters, and finally table calculation filters. This matters because, for example, a normal dimension filter does not affect a FIXED LOD unless that filter is added to context.
From the Analysis menu or the data pane, I create a calculated field with a formula, for example Profit Ratio = SUM([Profit]) / SUM([Sales]), or a category such as IF [Sales] > 10000 THEN "High" ELSE "Low" END. Calculations can be row-level or aggregate, but the two cannot be mixed in one expression without aggregating the row-level part.
A join combines tables at row level into one table before analysis, which can duplicate rows if the granularity differs.
Relationships (the default since Tableau 2020.2) keep tables separate and join them at the right level for each view, avoiding duplication.
Data blending combines data from different sources at an aggregated level using a primary and secondary source in the view.
Table calculations are computed on the results already in the view, such as running total, percent of total, difference from the previous period, moving average, rank and percentile.
Their result depends on how they are computed: across the table, down, or along specific dimensions, set in "Compute Using".
Functions include WINDOW_SUM, LOOKUP, RUNNING_SUM and INDEX.
Set actions let users change the members of a set by interacting with a visualisation, such as clicking or selecting marks. Combined with calculations that use the set, they allow advanced interactions such as highlighting selected categories while keeping others visible, drilling into one category at a time, or comparing a selected group against the rest.
Parameter actions update a parameter's value when a user selects or hovers on marks.
They make dashboards more dynamic: for example, clicking a date sets a parameter used to show details from that date, or selecting a value sets a reference line.
Combined with calculations, they enable features such as dynamic drill-down and custom comparisons.
A context filter is applied before other dimension filters, creating a temporary filtered dataset that later filters and top N calculations work from.
It is used to make a Top N filter respect another filter (such as top 10 products within a selected region) and to make FIXED LOD expressions respect a filter.
On large data sources, context filters can improve performance.
Quick table calculations add common calculations with one click on a measure: running total, difference, percent difference, percent of total, rank, percentile, moving average, year-to-date total, compound growth rate and year-over-year growth. They can then be edited to change how they are computed (across, down or by specific dimensions).
Filter actions filter other sheets based on a selection, highlight actions emphasise related marks, URL actions open web pages with values from the selected marks, go-to-sheet actions navigate to another sheet or dashboard, parameter and set actions change parameters and sets, and actions can run on hover, select or menu. They make dashboards interactive and connected.
Extract filters limit the data stored in an extract, such as only the last two years, which reduces extract size and speeds up performance.
Data source filters restrict data for every sheet using the data source, which is also useful for security when publishing.
Both apply before other filters, so users cannot see or select filtered-out data.
Tableau Prep Builder is a visual tool for cleaning and shaping data before analysis.
Flows connect to sources and apply steps such as cleaning values, splitting fields, pivoting, joining, unioning and aggregating, with a preview of the data at each step.
Flows output to extracts, published data sources or databases, and Prep Conductor schedules them on Tableau Server or Cloud.
System design, deep internals, leadership and tough follow-ups.
LOD expressions compute an aggregation at a level different from the view.
FIXED computes at the specified dimensions regardless of the view, such as {FIXED [Customer]: MIN([Order Date])} for first purchase date.
INCLUDE adds a dimension to the view's level, for example average sales per customer by region.
EXCLUDE removes a dimension, for example region total shown beside each state.
I use the Performance Recorder to find slow queries and rendering.
Then I use extracts and hide unused fields, filter data early with extract or data source filters, reduce the number of marks and sheets per dashboard, avoid many quick filters with "only relevant values", prefer boolean or integer calculations to string ones, and use context filters carefully.
On the database side, indexes and views help live connections.
I start with the questions users need answered, put the most important KPIs top left, and keep each dashboard focused.
I choose the right chart for each comparison, use consistent and colour-blind-friendly colours, add clear titles and tooltips, and use actions such as filter and highlight actions for interactivity.
I set a fixed or automatic size appropriate for the device and test with real users.
Common approaches are user filters, calculated fields using USERNAME() or ISMEMBEROF() compared with a field such as region, and an entitlements table joined to the data that maps users to the rows they may see, applied as a data source filter on a published data source. Virtual connections with data policies in Tableau Cloud centralise row-level security across workbooks.
Administrators organise content into sites and projects, manage users and groups with site roles (Creator, Explorer, Viewer), set permissions on projects with locked templates, schedule extract refreshes and subscriptions, monitor performance and usage with admin views, manage data connections and credentials, and on Server handle upgrades, backups and capacity.
The REST API automates administration tasks such as publishing, managing users, permissions and refreshes.
The Embedding API embeds interactive views in web applications with JavaScript and handles events and filters.
The Hyper API creates and updates .hyper extract files programmatically, the Metadata API queries lineage, and extensions add custom functionality to dashboards.
Tableau Pulse delivers automated metric insights to users in their workflow, such as Slack and email, explaining what changed and why in natural language. Tableau Agent (formerly Einstein Copilot for Tableau) helps users create calculations and visualisations by asking in plain language, and Explain Data suggests possible reasons for unexpected values.
Push work to the database with a live connection to a fast warehouse such as Snowflake or BigQuery, or use extracts aggregated to the needed level of detail with hidden unused fields and extract filters. Use initial SQL or custom SQL carefully, materialise heavy calculations in the data source, limit marks per view, and use incremental extract refresh for growing tables.
See all Data Scientist interview questions, covering Python, Pandas, NumPy, Statistics, Machine Learning and more.