Tree Maps

Understanding and using Tree Maps

The treemap functions as a visualization composed of nested rectangles. These rectangles represent certain categories within a selected dimension and are ordered in a hierarchy, or “tree.” Quantities and patterns can be compared and displayed in a limited chart space. Treemaps represent part to whole relationships.

Ben Shneiderman, a professor of computer science at the University of Maryland, developed this specific visualization to make the most of a small space. He wrote, “Tree structured node-link diagrams grew too large to be useful, so I explored ways to show a tree in a space-constrained layout.”

How do I read a Tree Map?

Treemaps provide an accessible way for viewers to interpret their data at a glance. Color can represent dimensions (such as categories) or measures (such as KPIs). If used to represent a KPI, a darker color may highlight extremes, high or low. For dimensions, a user might use a categorical palette, assigning a different color for each available shipping mode. For measures, a continuous color palette would show a company’s sales numbers or profit.

When analyzing a tree map for insights, the largest box shows the largest part of the whole, while the smallest box shows the smallest part. For a deeper analysis, These boxes can be nested to show many categories. For example, a box within the “office expenses” data set might display “percentage spent on furniture.” A box nested within that box might display “percentage spent on desks.”

What type of analysis does a tree map support?

Treemaps best depict data that needs to show a part-to-whole relationship. Percentages of a measure for each dimension are displayed as squares that, when added together, comprise the whole. Using our Superstore dataset, we can divide the total number of sales into categories. Adding the Product Sub-Category dimension would create “seating,” “tables,” and “lamps”. In this example, three specific items of furniture that fall under their respective categories become smaller nested squares.

In his visualization of data from the Electoral College in the 2012 U.S. Presidential election, Steve Wexler found a treemap was the best way to tell his data story. He used red and blue to distinguish between the Republican and Democratic parties, respectively, and each state’s vote for a candidate fell into either the red part or the blue part.

When and how to use Tree Maps for Visual Analysis

Treemaps are helpful when visualizing a large number of related categorie. If specific values should be prominently displayed, a bar chart may be a better choice.

A good treemap will have:

Distinct numerical values
A distinct hierarchy
No more than three or four labeled tiles
An obvious highest level of the hierarchy
Not display negative values

However, a poor treemap will:

Look cluttered with too many categories and labels
Have too many similar values

Great Examples of Tree Maps

This Tree Map looks at the listings per type of lodging available. The biggest box tells the viewer that most of the rooms available are the entire home or apartment, while the second biggest box measures the number of private rooms available.

This Tree Map only measures a few categories in relation to each other
The white borders indicate that all three measures make up a whole
The colors are distinct from each other

Ineffective examples of Tree Maps and Alternatives

POOR EXAMPLE

This treemap looks at the top ten hosts in a region for Airbnb and measures the categories by a number of listings. Unfortunately, the map is confusing and difficult to read.

The same color is used for multiple categories
Many of the boxes look similar in size and measure

BETTER ALTERNATIVE

A better alternative for an ineffective treemap would be a simple bar chart. This chart can be organized from largest to smallest, so the reader can easily compare to see who has the most listings and who has the fewest listings.