Tableau Public — Beginner Track

A full walkthrough of installing Tableau Public, connecting your first dataset, understanding the interface, and building your first bar chart, line chart, and simple dashboard.

1. What Tableau Public actually is

Tableau Public is a free version of Tableau's data visualization software. It works the same way the paid Tableau Desktop product does for building charts and dashboards, with two differences that matter: every workbook you save is published publicly to Tableau Public's gallery (there's no private/local save), and it can't connect to internal databases behind a company firewall — only flat files (Excel, CSV, text) and a limited set of public/cloud data sources. That trade-off is exactly why it's a good free lab environment: you get the real Tableau interface and the real drag-and-drop visual-building workflow with zero cost and zero install friction beyond the desktop app itself.

The beginner track gets you from "never opened Tableau" to "can take a spreadsheet and turn it into a clean chart and a two-visualization dashboard" without assuming any prior data-viz background.

2. Environment setup

1. Create a free account at public.tableau.com — this is required both to download the desktop app and to publish any workbook you build.
2. Download and install Tableau Public Desktop from the same site. It's a full desktop application, not a browser tool — installation takes a few minutes on either Windows or macOS.
3. Grab a starter dataset. Tableau Public ships with sample data (look for "Sample - Superstore" in the connection pane on first launch) — use that for this track rather than hunting for your own dataset, so you're learning the interface, not fighting messy data.
4. Open the app and connect to Sample - Superstore under "Saved Data Sources" or via Connect → To a File → Microsoft Excel if it's not pre-loaded.

3. The Tableau interface, in plain terms

Everything in Tableau revolves around dragging fields from a list on the left onto "shelves" that control what a chart shows. Learn these five areas before building anything:

4. Building your first chart

1. Double-click Category in the Data pane (a Dimension) — Tableau automatically adds it to Columns and switches to a bar-chart-friendly layout.
2. Double-click Sales (a Measure) — a bar chart appears immediately, one bar per category, showing the sum of Sales.
3. Drag Sub-Category onto the Rows shelf, after Category, to break each bar down further.
4. Drag Region onto Color on the Marks card — every bar now splits into colored segments by region.
5. Rename the sheet tab at the bottom (double-click it) to something descriptive like "Sales by Category".

Building a second chart type: a line chart over time

1. Open a new sheet (click the small sheet icon at the bottom, or right-click a tab → New Worksheet).
2. Double-click Order Date — Tableau adds it to Columns as a date hierarchy and automatically builds a line chart.
3. Double-click Profit to add it as the Y-axis measure.
4. Right-click the date field on Columns and try switching between Year, Quarter, and Month granularity to see how the line reshapes.

5. Filtering and sorting basics

• Filter: drag any field (e.g. Region) onto the Filters shelf, then choose which values to include or exclude — this restricts what the whole sheet displays.
• Sort: click the small sort icon that appears on the axis when you hover over a bar chart, or right-click the axis and choose Sort → Descending by the measure.
• Quick Filter (interactive): right-click a field already on the Filters shelf and choose "Show Filter" — this adds a clickable filter widget to the sheet itself, which becomes important once you build dashboards.

6. Formatting for clarity

• Use Label on the Marks card to show data values directly on bars — helpful for small charts where hovering for a tooltip is friction.
• Rename axis titles and legend titles by double-clicking directly on them in the view.
• Use Format → Font and Format → Borders from the top menu for consistent, presentation-ready styling rather than leaving Tableau's defaults.

7. Lab: build a two-chart dashboard

8. Pre-attempt checklist

• I know the difference between a Dimension (blue) and a Measure (green)
• I can build a bar chart and a line chart from scratch without following steps
• I can add a field to Color, Size, or Label on the Marks card
• I can add a filter and make it interactive with "Show Filter"
• I've published at least one workbook to Tableau Public

9. What to do after this lab

Compare your published dashboard against a few highly-viewed vizzes in the Tableau Public gallery (search "Superstore" — thousands of people have built on this same sample dataset, so it's a fair comparison). Note what makes the well-designed ones easier to read: consistent color use, clear titles, and restrained use of chart types. Once bar/line charts, filtering, and basic dashboards feel automatic, move to the Intermediate track, which covers calculated fields, table calculations, and multi-source joins.

← Back to all practice tools & labs

Tableau Public — Intermediate Track

Where drag-and-drop stops being enough. This guide covers calculated fields, table calculations, joining and blending multiple sources, parameters, and building a genuinely interactive dashboard.

1. What changes at this level

Beginner charts use fields exactly as they exist in your source data. Intermediate work is about creating new fields Tableau doesn't have out of the box — profit margin, a running total, a custom category grouping — and about combining more than one data source into a single view. This is also where dashboards stop being "two charts side by side" and start being genuinely interactive, with actions that let a user click one chart to filter or highlight another.

2. Calculated fields

A calculated field is a new column you define with a formula, built from existing fields. Create one via Analysis → Create Calculated Field, or right-click empty space in the Data pane → Create Calculated Field.

// Basic arithmetic between two existing measures
[Profit] / [Sales]

// A conditional (IF) calculation — bucket orders by size
IF [Sales] >= 1000 THEN "Large"
ELSEIF [Sales] >= 100 THEN "Medium"
ELSE "Small"
END

// String manipulation
LEFT([Customer Name], 1) + ". " + SPLIT([Customer Name], " ", -1)

// Date math
DATEDIFF('day', [Order Date], [Ship Date])

Common beginner-to-intermediate calculations worth practicing

• Profit Ratio: SUM([Profit]) / SUM([Sales]) — note the explicit SUM() here; without it, a calculation dividing raw row-level values instead of aggregates behaves differently once placed on a shelf.
• Days to Ship: DATEDIFF('day', [Order Date], [Ship Date])
• Custom grouping: an IF/ELSEIF chain bucketing a continuous measure into named tiers, as shown above.

3. Table calculations

Table calculations are different from calculated fields — they operate on the data after it's already been aggregated and arranged in the view, which makes them the right tool for running totals, rank, and percent-of-total. Right-click any measure already on a shelf and choose Quick Table Calculation, or build a custom one via Analysis → Create Calculated Field using functions like RUNNING_SUM, RANK, and WINDOW_AVG.

// Running total of Sales along the current table's partition
RUNNING_SUM(SUM([Sales]))

// Rank each Sub-Category by total Sales, highest = 1
RANK(SUM([Sales]))

// Percent of total — each mark's share of the whole view
SUM([Sales]) / TOTAL(SUM([Sales]))

4. Joins vs. blends — combining more than one data source

1. Practice a join first: connect Sample - Superstore's Orders and Returns tables, join them on Order ID, and build a chart showing Sales for only returned orders (drag Returned onto Filters).
2. Then practice a blend: connect a second, unrelated data source (even a small CSV you make yourself with a Region column) and blend it against Superstore on Region to combine values from both in one view.

5. Parameters — user-controlled inputs

A parameter is a single value a user of your dashboard can change, which then drives a calculated field, a filter, or a title. Create one via the dropdown in the Data pane → Create Parameter.

// A calculated field that swaps which measure is displayed,
// driven by a parameter named [Measure Selector]
CASE [Measure Selector]
    WHEN "Sales" THEN SUM([Sales])
    WHEN "Profit" THEN SUM([Profit])
    WHEN "Quantity" THEN SUM([Quantity])
END

Show the parameter control on your sheet (right-click the parameter in the Data pane → Show Parameter Control), and users can now switch what the chart displays without you building three separate charts.

6. Dashboard actions — real interactivity

Beyond the "Use as Filter" quick-toggle from the beginner track, Dashboard → Actions gives you fine control over three interaction types:

• Filter actions — clicking a mark in one sheet filters another. Configure which fields are passed and what happens when the selection is cleared.
• Highlight actions — clicking a mark highlights matching marks elsewhere without filtering anything out.
• Set actions / Parameter actions — clicking a mark updates a Set or Parameter value, which is how you build things like a chart that changes based on which bar a user last clicked.

7. Lab: an interactive, calculated-field-driven dashboard

8. Pre-attempt checklist

• I can write a calculated field using IF/ELSEIF logic and basic arithmetic
• I understand the difference between a calculated field and a table calculation
• I know when to join vs. blend two data sources
• I've built and shown at least one parameter control
• I've configured a filter action between two sheets on a dashboard

← Back to all practice tools & labs

Tableau Public — Expert Track

LOD expressions, advanced table calculations, performance optimization, and dashboard storytelling — enough to build something worth putting in a portfolio.

1. What changes at this level

Intermediate work solves "how do I calculate this" problems. Expert work solves "how do I calculate this at exactly the right level of granularity, keep it fast on a large dataset, and present it so a stranger understands the point in five seconds" — which is a design and performance problem as much as a formula problem. This track assumes you're comfortable with everything in the intermediate guide and are ready to work with a real dataset of your own choosing rather than Sample - Superstore.

2. Level of Detail (LOD) expressions

LOD expressions let you compute an aggregation at a granularity different from whatever's currently in the view — something regular calculated fields and table calculations can't do cleanly. There are three types:

// FIXED — always computes at the specified dimension(s), ignoring the view
{ FIXED [Customer Name] : SUM([Sales]) }
// → total sales per customer, no matter what else is on the sheet

// INCLUDE — adds a dimension to the computation beyond what's in the view
{ INCLUDE [Product Name] : SUM([Sales]) }
// → useful inside a view that's aggregated higher than product level

// EXCLUDE — removes a dimension from the computation
{ EXCLUDE [Sub-Category] : SUM([Sales]) }
// → e.g. total category sales, shown alongside sub-category rows

A worked example: customer segmentation with FIXED

// Step 1: total sales per customer, fixed regardless of view granularity
{ FIXED [Customer Name] : SUM([Sales]) }

// Step 2: bucket customers using that fixed value
IF { FIXED [Customer Name] : SUM([Sales]) } >= 5000 THEN "High Value"
ELSEIF { FIXED [Customer Name] : SUM([Sales]) } >= 1000 THEN "Mid Value"
ELSE "Low Value"
END

3. Advanced table calculations and addressing

At this level you should understand addressing and partitioning — which dimensions a table calculation "moves across" versus which it's "fixed within." Configure this explicitly via Edit Table Calculation → Compute Using → Specific Dimensions, ordering the dimensions to control the calculation direction precisely, rather than relying on Tableau's automatic default.

// A moving average table calc, explicit window
WINDOW_AVG(SUM([Sales]), -2, 0)   // average of current + prior 2 periods

// Year-over-year growth using LOOKUP
(SUM([Sales]) - LOOKUP(SUM([Sales]), -1)) / ABS(LOOKUP(SUM([Sales]), -1))

4. Performance optimization

Public workbooks with slow-loading vizzes get abandoned by viewers. A few habits that matter once your dataset grows past a few thousand rows:

• Use extracts, not live connections where possible — Tableau Public publishes an extract (a compressed snapshot of the data) rather than querying a live source on every view, which is both required for most Public data sources and generally faster.
• Hide unused fields in the Data pane (right-click → Hide) — this isn't just cosmetic, it reduces what Tableau has to account for when building the extract and the view.
• Avoid excessive quick filters on high-cardinality fields (e.g. a filter with thousands of unique Customer Names) — use a search-enabled filter or a parameter-driven approach instead.
• Check the Performance Recording tool (Help → Settings and Performance → Start Performance Recording) to see exactly which queries or rendering steps are slow before you guess at optimizations.

5. Dashboard layout and storytelling

Expert-level dashboards are built with intentional layout, not just sheets dropped onto a canvas:

• Switch from the default "Tiled" layout to explicit Floating containers when you need precise overlap (a KPI number floating over a background chart, for example).
• Use horizontal and vertical container objects to group related sheets so resizing the dashboard reflows predictably instead of breaking.
• Build a Story (a separate object type from a dashboard, found in the same tab bar) when you need to walk a viewer through a sequence of views with narrative captions — useful for portfolio pieces that need to explain themselves without you standing next to them.
• Set an explicit fixed-size dashboard (rather than "Automatic") when you know your target device, so text and marks don't rescale unpredictably.

6. Working with your own dataset

This track assumes you bring a real dataset rather than continuing with Sample - Superstore — public data from sources like data.gov, Kaggle, or a national statistics office works well. Before building anything:

1. Open the data source in Tableau and check every field's assigned data type (right-click a field header → Change Data Type) — misdetected dates or numbers-stored-as-text will quietly break calculations downstream.
2. Identify the actual grain of one row (one transaction? one day? one customer?) — this determines which LOD expressions you'll need and is the single most common source of double-counting bugs in dashboards built by people at this level.
3. Sketch the story you want the dashboard to tell before opening Show Me — expert vizzes are built backward from a question, not forward from whatever chart type looks good first.

7. Lab: a portfolio-ready, LOD-driven dashboard

8. Pre-attempt checklist

• I can explain the difference between FIXED, INCLUDE, and EXCLUDE LOD expressions
• I can set explicit Compute Using addressing on a table calculation rather than relying on the default
• I know how to identify the grain of a dataset before building anything
• I've used the Performance Recording tool at least once
• I've built a dashboard using container objects and at least one floating element

← Back to all practice tools & labs