Tableau Charts: How to Know Which Chart to Use, When to Use It, and Why It Matters
From simple bar charts to Sankey, Sunburst, Waffle, and other advanced visuals — a practical guide to turning data into a story.
When I first started working with Tableau, one of the things I found most interesting was the number of charts available.
Bar chart. Line chart. Pie chart. Tree map. Scatter plot. Heatmap. Maps. And then there were all the advanced visualizations that seemed to appear everywhere—Sankey diagrams, Sunburst charts, Waffle charts, Donut charts, and more.
At first, it was tempting to think:
“The more advanced the chart, the better the dashboard.”
But that is not really how good data visualization works.
A beautiful chart is not necessarily a useful chart.
The real skill in Tableau is knowing which chart will help your audience understand the data fastest.
For example, if your manager asks:
“Which region is performing best?”
you probably don't need a Sankey diagram.
A simple sorted bar chart may answer the question in seconds.
But if the question is:
“How are customers moving from website visits to purchases?”
Now a Sankey diagram starts making sense.
That is what this article is about.
Not just how to create Tableau charts, but how to decide which chart to use, why to use it, and when a particular chart should be avoided.
Before Choosing a Chart, Ask One Question
Before opening Tableau, ask yourself:
“What question am I trying to answer?”
This one question can save you from creating a dashboard full of attractive but confusing charts.
For example:
Business Question | A Good Starting Chart |
Which region has the highest sales? | Bar chart |
How are sales changing each month? | Line chart |
What percentage does each category contribute? | Donut / Pie |
Is discount related to profit? | Scatter plot |
Where are sales concentrated geographically? | Map |
Which months have the strongest sales? | Heatmap |
What does the sales distribution look like? | Histogram |
Are there unusual values? | Box plot |
How does a customer move through stages? | Sankey |
How is a hierarchy structured? | Sunburst |
How close are we to our target? | Bullet chart |
What percentage of a goal is complete? | Waffle chart |
This is the mindset I recommend developing as you learn Tableau.
Don't start with the visualization. Start with the question.
1. Bar Chart — Your Go-To Chart for Comparisons
If I had to choose one chart that every Tableau beginner should become comfortable with, it would be the bar chart.
Why?
Because comparison is one of the most common things we do with data.
Suppose your dataset contains sales for four regions:
East
West
Central
South
Your question is simple:
Which region generated the most sales?
A bar chart makes the answer immediately visible.
In Tableau
Drag:
Region → Rows
Sales → Columns
Then sort the bars from highest to lowest.
That's it.
You can make it more informative by adding labels or using color to highlight a specific category.
When should you use a bar chart?
Use it when you want to:
compare categories
rank products
identify Top N customers
compare regions
compare departments
show highest and lowest performers
When should you avoid it?
If you have dozens of categories, the chart can become difficult to read.
Instead, use a Top N filter, grouping, or another visualization.
A bar chart makes category-to-category comparisons easy to understand
Tableau Bar Chart – Sales by Region
2. Line Chart — When Time Becomes the Story
A line chart becomes useful when your question involves change over time.
Imagine your sales looked like this:
January → $80K
February → $85K
March → $92K
April → $110K
May → $105K
June → $125K
Looking at six numbers in a table is possible.
But a line makes the movement much easier to see.
You can immediately notice the increase, the small decline, and the later recovery.
Use a line chart for questions like:
Are sales growing?
Is revenue declining?
Which months have seasonal peaks?
Did performance improve after a campaign?
How has profit changed over several years?
In Tableau
Place:
Order Date → Columns
Sales → Rows
Then choose the appropriate date level—Month, Quarter, or Year.
One mistake I see often
People sometimes put too many categories on a line chart.
Suddenly you have 15 colored lines crossing each other.
Technically, the chart may be correct.
Practically, nobody wants to decode it.
Use filters, highlighting, or a smaller number of categories when necessary.
A line chart is ideal when the order of time and the movement of the metric matter.
Tableau Line Chart – Monthly Sales Trend
3. Stacked Bar Chart — When You Need the Total and Its Breakdown
Sometimes knowing the total isn't enough.
You also want to know what makes up that total.
Suppose East has $500K in sales.
You also want to know how much came from:
Technology
Furniture
Office Supplies
A stacked bar chart can show both pieces of information together.
In Tableau
Drag:
Region → Rows
Sales → Columns
Category → Color
Now each region becomes a bar divided into categories.
Best use case
Use a stacked bar when your audience needs to understand:
“How much?”
and
“What makes up that amount?”
Avoid it when...
You need to compare the individual middle segments very precisely.
In that situation, a grouped bar chart may be easier to interpre
4. Pie Chart — Keep It Simple
Pie charts are probably one of the most debated charts in data visualization.
I don't think pie charts are “bad.”
They simply have a specific job.
They work well when you're showing a small number of categories that together make one meaningful whole.
For example:
Technology — 45%
Furniture — 30%
Office Supplies — 25%
That's easy to understand.
But imagine having 15 categories.
Now the pie becomes a collection of tiny slices.
At that point, a bar chart will usually communicate the comparison much better.
Use a pie chart when:
there are only a few categories
percentages are the main message
all categories belong to one meaningful total
Don't use it when:
there are many categories
exact comparisons matter
you're showing a trend over time
Pie charts work best when a small number of categories form a meaningful whole.
Tableau Pie Chart – Sales Contribution by Category
5. Donut Chart — Same Idea, Different Presentation
A donut chart is basically a pie chart with a hole in the middle.
So why use one?
The center gives you an opportunity to display an important number.
For example:
$2.4M
Total Sales
Then the surrounding segments show how that total is distributed.
That makes a donut especially useful in dashboards where you want a visual plus a central KPI.
Good examples
Sales mix
Expense distribution
Customer segments
Completion percentage
Revenue contribution
One important point
Don't create a donut chart just because it looks modern.
If the audience needs to compare six categories accurately, a bar chart may still be better.
6. Scatter Plot — When You Want to Ask “Are These Two Things Related?”
This is one of my favorite charts for analysis.
Suppose you have:
Discount
and
Profit
You may wonder:
“Are higher discounts associated with lower profit?”
A scatter plot lets you investigate that relationship.
Each dot represents an observation such as a customer, product, or order.
You may see:
an upward pattern
a downward pattern
no obvious relationship
clusters
unusual outliers
You can also add a trend line to help understand the overall relationship.
Use scatter plots for:
Sales vs Profit
Advertising Spend vs Revenue
Age vs Income
Price vs Quantity
Discount vs Profit
Important reminder
A relationship in a scatter plot does not automatically prove causation.
The chart helps you investigate a relationship. It doesn't automatically explain why the relationship exists.
Scatter plots help analysts explore relationships, clusters, and potential outliers.
Tableau Scatter Plot – Sales vs Profit
7. Histogram — What Does the Data Actually Look Like?
Sometimes the average doesn't tell you enough.
Imagine the average delivery time is 5 days.
That sounds reasonable.
But what if:
most orders arrive in 2–3 days
a smaller group takes 10–15 days
a few orders take 30 days?
The average hides that story.
A histogram helps you see the distribution.
Use it for:
Age
Salary
Order value
Delivery time
Customer purchase amount
Response time
The important idea is that the values are grouped into ranges, or bins.
8. Box Plot — Quickly Spot Spread and Outliers
Box plots are especially useful when you want to compare distributions between groups.
Suppose you're comparing delivery time across four regions.
A box plot can help you understand:
the median
the spread
the range
potential outliers
This is much more informative than simply comparing the average delivery time.
Use a box plot when your question is:
“How consistent is this metric?”
rather than simply:
“What is the average?”
9. Heatmap — Let Color Reveal the Pattern
Heatmaps are fantastic when you have two dimensions and want to find patterns quickly.
For example:
Region × Month
with Sales represented by color.
You may immediately notice that:
one region performs strongly in December
another region is consistently weak
certain months perform well across all regions
That's information you might miss in a simple table.
Great use cases
Sales by month and region
Website activity by day and hour
Employee performance
Product performance
Customer activity
Heatmaps use color intensity to make patterns and high/low areas easier to spot.
Tableau Heatmap – Sales by Region and Month
10. Highlight Table — When You Need Both Color and Numbers
A highlight table takes the idea of a heatmap one step further.
You get the color pattern and the actual number.
This is useful when your audience wants to quickly identify patterns but still needs exact values.
For example:
Region | January | February | March
The color tells you which cells are performing strongly, while the numbers tell you exactly how much.
This is a good choice for analytical dashboards where precision matters.
11. Treemap — When Hierarchy and Size Matter
Treemaps are great when you want to show hierarchical categories using size.
Imagine:
Technology
→ Phones
→ Computers
→ Accessories
The size of each rectangle can represent Sales.
The larger the rectangle, the larger the value.
Use treemaps when:
you have hierarchical categories
you want to show part-to-whole relationships
you have many categories
space on the dashboard is limited
A treemap can communicate a large amount of information in a relatively small area.
12. Maps — Use Them When Location Is Part of the Question
Having a geographic field does not automatically mean you need a map.
This is an important distinction.
If your question is:
“Which state has the highest sales?”
a sorted bar chart may actually be better.
But if your question is:
“Where are our sales concentrated?”
then a map becomes much more useful.
Use maps for:
sales by state
customers by city
store locations
patient locations
delivery locations
geographic trends
The map should add geographic meaning—not just decoration.
Maps are most useful when geographic location is part of the business question.Tableau Filled Map – Sales by State
13. Bullet Chart — Actual vs Target
If you work with KPIs, you will probably use bullet charts often.
Imagine:
Sales Target: $1M
Actual Sales: $850K
A bullet chart allows the viewer to understand the performance against the target in one compact visual.
Use it for:
Sales vs Target
Actual vs Budget
Revenue vs Goal
Employee performance
Service-level targets
It is particularly useful when dashboard space is limited.
14. Gantt Chart — When Duration Matters
Gantt charts are about time duration.
They can answer questions such as:
When did the project start?
How long did it take?
Which tasks overlap?
Which stage is taking the longest?
They work well for project management, scheduling, operations, and timelines.
15. Waterfall Chart — Tell the Story Behind a Change
A waterfall chart is excellent for explaining how you moved from one number to another.
Imagine your profit started at:
$500K
Then:
$100K New Customers
$50K Upselling
− $70K Discounts
− $40K Returns
Final Profit:
$540K
Instead of just showing $500K and $540K, the waterfall explains what happened in between.
This makes it particularly useful for finance and business reporting.
Packed Bubble Charts: Packed Bubble Charts use the size of each bubble to represent a measure, making it easy to spot the largest and smallest categories at a glance. They work well for sales by category, customer revenue, product performance, or market share. Use them when you want a quick visual comparison rather than precise value-by-value analysis. For exact comparisons, a bar chart is usually a better choice.
Ask yourself: “Do I want my audience to quickly see which categories are bigger or smaller?” If yes, a Packed Bubble Chart can be a good choice.
Now Let's Talk About Advanced Tableau Charts
Once you are comfortable with the basic charts, you can start experimenting with advanced visualizations.
But here's my favorite rule:
Don't use an advanced chart to impress your audience. Use it to explain something that a basic chart cannot explain as effectively.
Let's look at some of the most useful ones.
16. Sankey Diagram — Perfect for Showing Flow
A Sankey chart becomes useful when something moves from one stage to another.
Think about a customer journey:
Website
↓
Product Page
↓
Add to Cart
↓
Checkout
↓
Purchase
The width of the flow can represent the number of customers.
Now you can see where customers are moving and where they are dropping off.
Other examples
Sankey charts can be used for:
customer journeys
recruitment pipelines
patient journeys
supply chains
budget allocation
energy flow
product movement
Ask yourself:
“Is something moving from A to B?”
If yes, a Sankey may be worth considering.
Sankey diagrams are useful when the movement or flow between stages is the main story.
Sankey Diagram – Customer Journey
17. Sunburst Chart — Show Multiple Levels of a Hierarchy
Sunburst charts are visually interesting because they display hierarchy through multiple rings.
For example:
Category
→ Sub-Category
→ Product
The center represents the highest level, while the outer rings represent deeper levels.
Use it when:
hierarchy is important
you have multiple levels
you want to explore parent-child relationships
Example
Center:
Technology
Next ring:
Phones | Computers | Accessories
Outer ring:
Individual products
The result gives the viewer a sense of how the smaller pieces fit into the larger structure.
But remember...
If you only have one or two categories, don't use a Sunburst.
A bar chart may tell the story much more clearly.
18. Waffle Chart — A Simple Way to Show Percentages
A Waffle chart uses a grid—often 100 small squares—to represent a percentage.
Suppose:
82% of customers are satisfied.
You can highlight approximately 82 of the 100 squares.
It is visually simple and works particularly well in executive dashboards.
Good use cases
Completion rate
Satisfaction rate
Achievement percentage
Conversion rate
Goal progress
One thing I recommend is displaying the exact percentage as a large number as well.
The visual gives the impression.
The number gives the precision.
Waffle charts work well when the main message is a percentage out of 100
Tableau Waffle Chart – 82% Customer Satisfaction
19. Onion Chart — For Layered Information
The Onion chart is a more specialized visualization.
It can be useful when your data naturally has multiple layers surrounding a central concept.
For example:
Overall Performance
→ Region
→ Category
→ Sub-Category
→ Product
The layered design can help communicate relationships between levels.
However, this is one of those charts where I would ask:
“Does the visualization actually make the hierarchy easier to understand?”
If the answer is no, use a treemap, bar chart, or another simpler option.
Advanced does not always mean better.
20. Dual-Axis Chart — Two Measures, One Story
Sometimes you want to look at two related metrics together.
For example:
Sales
and
Profit Margin
You might use:
Bars → Sales
Line → Profit Margin
This creates a dual-axis visualization.
Good use cases
Sales + Profit Margin
Revenue + Growth %
Actual + Target
Orders + Average Order Value
Be careful
Dual-axis charts can become misleading if the scales are poorly designed.
Always make sure the viewer understands what each axis represents.
21. Combination Chart — When One Chart Isn't Enough
A combination chart brings different visual forms together.
For example:
Bars = Revenue
Line = Growth Rate
This can be useful when one measure gives you the size of the business while another shows how quickly it is changing.
It is especially useful in executive dashboards.
What About Tableau's Show Me?
If you are still learning Tableau and aren't sure which visualization to choose, Show Me is a great starting point.
Select the fields you want to analyze and open Show Me.
Tableau will suggest visualization types based on the fields you selected.
This is a good way to learn the relationship between:
Data type → Analytical question → Visualization
Over time, you won't need Show Me as much because you'll begin recognizing the patterns yourself.
Tableau's Show Me panel helps beginners understand which visualizations can be created from selected fields.
Tableau Show Me Panel
A Real Example: Choosing Charts for a Sales Dashboard
Let's say you're building a dashboard using a sales dataset.
You have:
Order Date
Region
Category
Product
Sales
Profit
Discount
Customer
State
Now let's decide which chart should answer each question.
“How much are we selling?”
KPI Card
“How are sales changing?”
Line Chart
“Which region performs best?”
Sorted Bar Chart
“Which categories contribute to sales?”
Treemap or Donut
“Does discount affect profit?”
Scatter Plot
“Which states have the strongest sales?”
Map
“Which months and regions perform best?”
Heatmap
“How is profit distributed?”
Histogram / Box Plot
“How do customers move through the buying process?”
Sankey
Now your dashboard isn't just a collection of charts.
Every visualization has a job.
And that is the difference between a dashboard that looks good and a dashboard that actually helps people make decisions.
My Simple Chart Selection Formula
Whenever you're confused about which chart to use, try this:
If you're asking “Which is bigger?”
→ Bar Chart
If you're asking “How is it changing?”
→ Line Chart
If you're asking “Are these related?”
→ Scatter Plot
If you're asking “What makes up the total?”
→ Donut / Pie / Treemap
If you're asking “Where is it happening?”
→ Map
If you're asking “Where are the patterns?”
→ Heatmap
If you're asking “How is the data distributed?”
→ Histogram / Box Plot
If you're asking “How close are we to the target?”
→ Bullet Chart
If you're asking “How did we get from A to B?”
→ Waterfall
If you're asking “How does something move?”
→ Sankey
If you're asking “How is the hierarchy structured?”
→ Sunburst / Treemap
If you're asking “What percentage is complete?”
→ Waffle
One Last Piece of Advice
When you're creating Tableau dashboards, don't try to use every chart you know.
I've seen dashboards where almost every available visualization has been placed on one screen.
There may be:
A pie chart.
A donut.
A treemap.
A Sankey.
A map.
Three line charts.
Four KPI cards.
A scatter plot.
And a few filters.
It may look impressive in a screenshot.
But the viewer has no idea where to look first.
A good dashboard should have visual hierarchy.
Start with the most important question.
Then provide supporting analysis.
For example:
KPI → What happened?
Line chart → When did it happen?
Bar chart → Where did it happen?
Scatter plot → What relationships might explain it?
Advanced chart → How does the process or hierarchy work?
That creates a story.
Conclusion:
Tableau gives us a huge number of ways to visualize data, but you don't need to use all of them.
The real skill is not knowing how to create a Sankey diagram.
The real skill is knowing when a Sankey diagram is actually necessary.
Start simple.
Understand your data.
Understand your audience.
Identify the question.
Then choose the visualization.
And if a simple bar chart answers the question better than a complicated custom visualization, choose the bar chart.
Because at the end of the day, the goal of a Tableau dashboard isn't to make people say:
“Wow, that's a fancy chart.”
The goal is to make them say:
“Now I understand what is happening.”
And that, in my opinion, is what good data visualization is really about.
Thank you so much for taking the time to read my blog! I truly appreciate your support and interest in my content.
