Size encoding in Tableau - Deep Dive

Overview - Size encoding

What is it?

Size encoding is a way to show data by changing the size of marks or shapes in a visualization. Bigger marks mean bigger values, and smaller marks mean smaller values. This helps people quickly see differences in data amounts or importance. It is often used in charts like scatter plots or maps.

Why it matters

Size encoding helps people understand data quickly without reading numbers. Without it, viewers might miss important differences or patterns because all marks look the same. It makes data stories clearer and faster to grasp, which is important for making good decisions.

Where it fits

Before learning size encoding, you should know basic chart types and how to add data fields to visuals. After mastering size encoding, you can learn about color encoding and advanced visual techniques like combined encodings or interactive dashboards.

Mental Model

Core Idea

Size encoding uses the size of visual marks to represent the magnitude of data values, making differences easy to see at a glance.

Think of it like...

It's like using different sized boxes to pack items: bigger boxes hold more stuff, smaller boxes hold less, so just by looking at box sizes you know how much each holds.

Data Value → Size of Mark

┌─────────────┐
│   Small     │  ← Small value → Small mark
└─────────────┘

┌───────────────┐
│    Medium     │  ← Medium value → Medium mark
└───────────────┘

┌───────────────────┐
│      Large        │  ← Large value → Large mark
└───────────────────┘

Build-Up - 6 Steps

1

Foundation

What is size encoding in Tableau

Concept: Introduce size encoding as changing mark size to show data values.

In Tableau, size encoding means adjusting the size of marks (like circles or bars) based on a data field. For example, sales numbers can control how big each circle is on a map. You drag a measure to the Size shelf to do this.

Result

Marks on the visualization change size according to the data values you chose.

Understanding that size can represent data magnitude helps you add a new visual dimension to your charts.

2

Foundation

How to apply size encoding in Tableau

Concept: Learn the steps to add size encoding to a visualization.

1. Create a chart with marks (e.g., scatter plot). 2. Drag a numeric field (measure) to the Size shelf on the Marks card. 3. Tableau automatically adjusts mark sizes based on that field. 4. Use the Size slider to fine-tune the size range.

Result

The chart now shows marks sized by the data values you selected.

Knowing the exact steps lets you quickly add size encoding to any chart.

3

Intermediate

Choosing the right data for size encoding

🤔Before reading on: do you think size encoding works best with categorical or continuous data? Commit to your answer.

Concept: Size encoding works best with continuous numeric data to show magnitude differences.

Size encoding is most effective when used with continuous measures like sales, population, or revenue. Using categorical data (like product names) won't work well because size needs numeric values to scale marks properly.

Result

Using continuous data for size encoding makes differences clear and meaningful.

Understanding data types prevents confusing or misleading visualizations.

4

Intermediate

Balancing size encoding with other visual elements

🤔Before reading on: do you think making all marks very large improves clarity or causes clutter? Commit to your answer.

Concept: Size encoding must be balanced with other visual elements to avoid clutter or confusion.

If marks become too large, they can overlap and hide information. Use Tableau's Size slider to adjust the maximum and minimum sizes. Combine size encoding with color or shape to add more data dimensions without overwhelming the viewer.

Result

A clear, balanced visualization that uses size effectively without clutter.

Knowing how to balance size with other visuals improves readability and user experience.

5

Advanced

Interpreting size encoding with perception limits

🤔Before reading on: do you think doubling a data value always means doubling the perceived size? Commit to your answer.

Concept: Human perception of size is not linear; doubling data values doesn't double perceived size.

People perceive area (size) differently than raw numbers. For example, a circle with twice the diameter has four times the area, which can exaggerate differences. Tableau uses area scaling for size encoding, so be careful interpreting exact values from size alone.

Result

Better understanding of how size encoding affects viewer perception and potential misinterpretations.

Knowing perception limits helps you design more accurate and honest visualizations.

6

Expert

Advanced size encoding with combined encodings

🤔Before reading on: do you think size encoding can replace color encoding in all cases? Commit to your answer.

Concept: Combining size encoding with color or shape can reveal complex data stories but requires careful design.

In advanced dashboards, size encoding is often combined with color encoding and shape to show multiple data dimensions simultaneously. For example, size can show sales volume, color can show profit margin, and shape can show product category. This layered approach requires balancing to avoid confusion.

Result

Rich, multi-dimensional visualizations that communicate complex insights effectively.

Understanding how to combine encodings unlocks powerful storytelling in data visualization.

Under the Hood

Tableau maps data values to mark sizes by calculating relative proportions. It uses area scaling, meaning the area of a mark changes proportionally to the data value, not just the length or diameter. This is because humans perceive area better than length for size differences. Tableau normalizes data values between minimum and maximum sizes set by the user, then adjusts mark sizes accordingly.

Why designed this way?

Area scaling was chosen because it aligns better with human perception, making size differences more intuitive. Linear scaling by length or height would mislead viewers by underrepresenting differences. Tableau's design balances visual clarity with perceptual accuracy, avoiding misleading charts while keeping them easy to read.

Data Value → Normalization → Area Scaling → Mark Size

┌─────────────┐
│ Raw Data    │
└──────┬──────┘
       │
       ▼
┌─────────────┐
│ Normalize   │  (Scale values between min and max)
└──────┬──────┘
       │
       ▼
┌─────────────┐
│ Area Scale  │  (Calculate mark area proportional to value)
└──────┬──────┘
       │
       ▼
┌─────────────┐
│ Render Mark │  (Draw mark with calculated size)
└─────────────┘

Myth Busters - 4 Common Misconceptions

Quick: Does a mark twice as big in diameter represent twice the data value? Commit to yes or no.

Common Belief: A mark twice as big in diameter means twice the data value.

Reality: Because size encoding uses area, a mark twice the diameter actually represents four times the data value.

Why it matters: Misunderstanding this leads to overestimating differences and misreading the data story.

Quick: Can size encoding be used effectively with categorical data? Commit to yes or no.

Common Belief: Size encoding works well with any data type, including categories.

Reality: Size encoding requires numeric continuous data to scale marks properly; categorical data cannot be sized meaningfully.

Why it matters: Using categorical data for size causes confusing visuals that don't communicate value differences.

Quick: Does increasing mark size always improve chart clarity? Commit to yes or no.

Common Belief: Making marks larger always makes the chart easier to understand.

Reality: Too large marks cause overlap and clutter, reducing clarity and hiding data points.

Why it matters: Ignoring this leads to messy charts that confuse viewers instead of helping them.

Quick: Is size encoding the best way to show all types of data differences? Commit to yes or no.

Common Belief: Size encoding is the best way to show all data differences.

Reality: Size encoding is best for magnitude differences but not for showing categories or exact values; other encodings like color or labels may be better.

Why it matters: Relying only on size encoding limits the effectiveness of your visualization and may mislead users.

Expert Zone

1

Size encoding uses area scaling, but the exact scaling curve can be adjusted in Tableau to emphasize or de-emphasize differences.

2

Combining size encoding with interactivity (like tooltips or filters) helps overcome perception limits by showing exact values on demand.

3

In dense visualizations, subtle size differences may be lost; experts often combine size with other encodings or use aggregation to maintain clarity.

When NOT to use

Avoid size encoding when data values are very close or when precise comparisons are needed; use labels or color gradients instead. Also, do not use size encoding for categorical data or when marks overlap heavily, as it reduces clarity.

Production Patterns

Professionals use size encoding in dashboards to highlight key metrics like sales volume on maps or scatter plots. They combine it with color to show profitability and use filters to let users explore data subsets. Size encoding is often paired with legends and tooltips to aid interpretation.

Connections

Color encoding

Complementary visual encoding

Knowing how size and color encoding work together helps create richer, multi-dimensional visualizations that communicate more information clearly.

Human perception psychology

Underlying principle

Understanding how humans perceive area versus length explains why size encoding uses area scaling, improving visualization design.

Cartography

Shared visualization technique

Size encoding is similar to how map makers use symbol sizes to represent population or traffic, showing how BI visualization borrows from map design principles.

Common Pitfalls

#1 Using size encoding with categorical data causing meaningless sizes.

Wrong approach:

Drag a category field like 'Product Type' to Size shelf to size marks.

Correct approach:

Use a numeric measure like 'Sales' on the Size shelf to size marks meaningfully.

Root cause:

Misunderstanding that size encoding requires numeric continuous data to scale marks properly.

#2 Setting all marks to very large size causing overlap and clutter.

Wrong approach:

Set Size slider to maximum for all marks without adjustment.

Correct approach:

Adjust Size slider to balance visibility and avoid overlap, keeping marks readable.

Root cause:

Not considering visual clutter and mark overlap when increasing sizes.

#3 Assuming size differences represent linear data differences.

Wrong approach:

Interpreting a mark twice the diameter as twice the data value.

Correct approach:

Understand that size encoding uses area scaling, so diameter differences represent squared data differences.

Root cause:

Ignoring human perception principles and area scaling in size encoding.

Key Takeaways

Size encoding changes the size of marks to represent data values, making differences easy to see.

It works best with continuous numeric data, not categories, to scale marks properly.

Human perception of size is based on area, so size encoding uses area scaling, not linear scaling.

Balancing size with other visual elements prevents clutter and improves clarity.

Combining size encoding with color and interactivity creates powerful, multi-dimensional visualizations.