Building Views and Core Chart Types

You connected your data. Now you open a new worksheet, and the screen is almost empty. On the left is a list of your fields. In the middle is a wide blank canvas. Your manager just asked a simple question: “Which region sold the most last quarter?” The answer is sitting in that data. Your job is to turn it into a picture anyone can read in three seconds.

That picture is called a view in Tableau. This chapter shows you how to build one from a blank worksheet, and how to pick the right kind of chart for the question you were actually asked.

Why this matters

Most people who open Tableau for the first time feel stuck at the blank worksheet. They have the data. They just do not know where to put it.

The good news is that Tableau works by dragging and dropping. You do not write code. You take a field, drop it in the right spot, and a chart appears. Once you learn where the “right spots” are, you can answer a business question in under a minute.

The other half of the skill is judgment. Tableau will happily build you a chart that looks fine and answers the wrong question. A pie chart with fourteen slices technically works, but nobody can read it. Knowing which chart fits which question is what separates a useful dashboard from a confusing one. This chapter covers both.

The drag-and-drop model: dimensions, measures, and shelves

Before you build anything, you need two words.

A dimension is a field that describes or categorizes. Region, Product Name, Order Date, Customer, and Country are dimensions. They are usually the labels you group by. In the Data pane on the left, dimensions typically appear in blue.

A measure is a field you can add up or do math on. Sales, Profit, Quantity, and Discount are measures. They are usually numbers you want to total or average. Measures typically appear in green.

At the top of the worksheet are two long strips called shelves: the Columns shelf and the Rows shelf. Whatever you drop on Columns spreads left to right across the screen. Whatever you drop on Rows stacks top to bottom. That is the whole trick.

Here is the key behavior to remember. When you drop a measure into your view, Tableau does not show every single row of data. It aggregates the measure, which means it rolls the numbers up into one total per category. By default that total is usually a SUM. So if you drop Sales next to Region, Tableau shows the sum of sales for each region, written as SUM(Sales), not thousands of individual sale amounts.

You can change that later (to average, count, or others), but knowing the default is SUM explains most of what you see on screen early on.

A quick mental picture of a finished bar chart:

Columns:  [ Region ]
Rows:     [ SUM(Sales) ]

Sales
  |   ####
  |   ####        ####
  |   ####   ##   ####   ##
  +--------------------------
     West  East  South Central

The Marks card: how your chart looks

To the left of the canvas, below the shelves, is the Marks card. A “mark” is a single shape in your chart: one bar, one dot, one line segment, one square. The Marks card controls how those shapes look and what extra information they carry.

At the top of the Marks card is the mark type dropdown. It usually reads Automatic, which lets Tableau pick the shape for you. You can override it and choose a specific type. The available mark types are Automatic, Bar, Line, Area, Square, Circle, Shape, Text, Map, Pie, Gantt Bar, Polygon, and Density.

Below the dropdown are property buttons. You add detail to a chart by dragging a field onto one of these:

Color paints your marks by a field. Drag Region onto Color and each region gets its own color.
Size makes marks bigger or smaller based on a number. Bigger dot means bigger value.
Label prints the value or name directly on the mark, so people do not have to guess.
Detail breaks marks apart by a field without changing color or size. It adds granularity quietly.
Tooltip controls the little pop-up box that appears when someone hovers over a mark.

Color, Label, Detail, and Tooltip can each hold more than one field. Size and Shape hold one field at a time.

The Marks card is where a plain chart becomes an informative one. It is also where beginners overdo it, which we will get to in the pitfalls.

Show Me: a starting point, not the boss

In the top-right corner of the toolbar is a button called Show Me. Click it and a panel opens with thumbnail pictures of chart types.

Show Me looks at the fields you have selected and lights up only the chart types those fields can support. If you hover over a thumbnail, it tells you the minimum fields that chart needs. Tableau also marks its top suggestion with an orange outline. For example, if you select a date dimension and a measure, Show Me will suggest a line chart.

A fast way to use it: in the Data pane, hold Ctrl (Command on a Mac) and click the two or three fields you care about, then click Show Me and pick a chart. Starting in version 25.2, Show Me can even build a chart from a nearly empty worksheet using a “Choose for me” option.

Here is the judgment part. Show Me is a good launcher when you are exploring and do not know what fits. Trust it to get a first draft fast. But it does not know your business question. It picks based on field types, not on what your audience needs to understand. Once you know the question, build the view by hand so you stay in control. Use Show Me to explore; override it to communicate.

Try It: build a bar chart, then a line chart

Open a new worksheet with any dataset that has a category field, a date field, and a sales-type number (Tableau’s built-in Sample - Superstore works well).

Build a bar chart:

From the Data pane, drag Region (a dimension) to the Columns shelf.

Drag Sales (a measure) to the Rows shelf. Tableau draws one bar per region and labels it SUM(Sales).

On the Marks card, confirm the mark type dropdown reads Automatic or Bar.

Drag Region onto the Label button on the Marks card so each bar prints its name.

Now build a line chart from the same data:

Click the worksheet tab at the bottom to open a fresh sheet.

Drag Order Date to Columns. Tableau treats dates as time and spreads them left to right.

Drag Sales to Rows. You now have a line showing sales over time.

Right-click the date on Columns and switch between Year, Quarter, and Month to see the trend at different zoom levels.

What did you notice? The same two-field move produced two different chart types, because Tableau read Region as a category and Order Date as time.

Core chart types and the question each one answers

The mechanics are easy once you know the shelves. The real skill is matching the chart to the question. Here are the workhorses.

Bar chart: for comparing amounts across categories. “Which region sold the most?” “Which product returned the least?” Bars line up side by side and the eye compares length instantly. This is the safest default for comparison questions. When in doubt, use a bar.

Line chart: for trends over time. “Are sales growing or shrinking?” “When does demand spike?” A date on Columns and a measure on Rows gives you a line. Lines are the right choice whenever time is involved and you care about the direction of change.

Map: for questions about place. “Which states are underperforming?” Tableau can plot geography automatically. When a field holds a recognized place name like Country, State, or City, Tableau assigns it a geographic role, which links each value to a latitude and longitude. It also creates two hidden fields, Latitude (generated) and Longitude (generated). Drag a geographic dimension onto Detail on the Marks card, or double-click it, and Tableau draws a map. If a location field is not recognized, you can assign the role yourself in the Data pane by clicking the field’s data type icon, choosing Geographic Role, and picking the right one.

Scatter plot: for relationships between two numbers. “Do bigger discounts actually cost us profit?” Put one measure on Columns and another on Rows. Each dot is one item, positioned by both numbers at once. Scatter plots reveal whether two things move together.

Crosstab (text table): for the exact numbers. Sometimes people want to read values, not shapes. A crosstab is a grid of numbers, like a spreadsheet inside Tableau. Build it by putting dimensions on Rows and Columns and dragging a measure onto Text on the Marks card. Use it when precision matters more than pattern, such as a finance table someone will audit.

Highlight table: a crosstab with color. Take that grid of numbers and add color shading so high and low values pop out. Drag your measure onto Color as well as Text. Now the reader sees the exact figures and the pattern at the same time. It is a good middle ground when people want both.

The pattern to internalize: a comparison question wants a bar, a time question wants a line, a place question wants a map, a relationship question wants a scatter, and a “show me the numbers” question wants a table.

Sorting and light formatting

Two small moves make a chart much more readable.

Sorting puts your bars in order so the biggest or smallest jumps out. Hover over the axis of a bar chart and a small sort icon appears; click it to sort ascending or descending. A sorted bar chart answers “which is highest” without the reader doing any work.

Formatting is about clarity, not decoration. Right-click a number axis to format it as currency or to drop unnecessary decimal places. Right-click labels to change font size. The goal is always readability. If a formatting change does not help someone understand faster, skip it.

A gentle look at dual-axis (combination) charts

Sometimes one question needs two measures with different scales. “How did sales and profit ratio move together this year?” Sales might be in the thousands and profit ratio in single-digit percentages. Plotting them on one axis would flatten one of them.

A dual-axis chart solves this by giving each measure its own vertical axis. Here is the gentle version:

Drag your first measure to Rows and build a normal chart over time.
Drag your second measure to Rows as well, so it sits beside the first.
Right-click the second measure on Rows and choose Dual Axis. Tableau layers the two charts on top of each other, each with its own axis.

Because each measure now has its own Marks card, you can make one a bar and the other a line, which is why these are also called combination charts. One warning: if the two measures share the same units, right-click the right-hand axis and choose Synchronize Axis so both use the same scale. Skipping that step is a common source of misleading charts.

Keep dual-axis in your back pocket. Reach for it only when a single question genuinely needs two measures compared over the same time frame.

In business

A regional sales lead at a mid-size distributor gets the same request every month: a one-page read on how the territories are doing. In the past she exported a spreadsheet and pasted numbers into an email that nobody read closely.

Now she builds three quick views. A bar chart of SUM(Sales) by Region, sorted descending, makes the strongest and weakest territories obvious at a glance. A line chart of Sales by Order Date at the month level shows whether the recent dip is a trend or a blip. A map of Sales by State reveals that two neighboring states are dragging down an otherwise healthy region.

Each view answers one question: who is ahead, which way are we heading, and where is the problem. Together they take her ten minutes to build and give her manager a clear picture without a paragraph of explanation. In the next chapter you will combine views like these onto a single dashboard.

Common pitfalls

Choosing the wrong chart for the question. The most frequent mistake is a mismatch, like using a pie chart to compare fourteen products (bars are far easier to read) or a bar chart to show a trend over three years (a line shows direction better). Before you build, say the question out loud, then pick the chart that answers exactly that. Do not let the tool decide for you.

Overusing color. Color is powerful because the eye reads it first. That is also why it is easy to abuse. Dropping a field on Color that already appears on an axis adds noise without adding meaning. A rainbow chart looks busy and says little. Use color to encode one thing the reader needs, such as profit or category, and leave it off when the axis already tells the story.

Dual-axis confusion. Two axes on one chart can mislead as easily as they inform. If the two axes use different scales and you do not label them clearly, a reader may think two lines cross in a meaningful way when they do not. And if the measures share units but you forget to Synchronize Axis, the chart quietly lies about the comparison. Use dual-axis sparingly, label both axes, and synchronize when the units match.

Practice

Open Sample - Superstore or your own dataset. Build a sorted bar chart of a measure by a category, then add the category to Label. Write down the one sentence it answers.
Build a line chart of the same measure over time. Switch the date between Year, Quarter, and Month. Note which level makes the trend clearest.
Build a map: find a geographic field, confirm it has a geographic role, and drag it into the view. Add a measure to Color to shade the map.
Take any chart you built and open Show Me. Look at what Tableau suggests with the orange outline. Decide whether its suggestion answers your question better than your original choice, and explain why in one sentence.
For a business question of your own (“which client is most profitable,” “is our churn rising,” “where are our customers”), name the chart type you would use before opening Tableau. Then build it and check whether your first instinct was right.

Next, Chapter 5 shows how to assemble several of these views into a single interactive dashboard. Chapter 6 introduces calculations, so you can build fields your data does not contain yet, like profit ratio or year-over-year growth.

Part of the Tableau track. Developed with AI assistance and reviewed by a human editor. Tableau changes often; menu names, features, and prices may have shifted since the “checked on” dates in this chapter. Verify anything critical against Tableau’s official documentation.