EDGE-EQUALIZED TREEMAPS - Information Visualisation: Techniques, Usability and Evaluation - page 109

personal achievements in an organizational structure could be used as
weights.
A Treemap is a visualization method often used to represent large sets
of hierarchical data. It is easy to understand small sets of hierarchical data,
but this is not always so for large hierarchical datasets. We often use
visualization techniques to understand such large data. In Treemaps [1],
[3], [5], [6], [8], nodes are drawn as rectangles and arranged in a nested
style (we discuss the detail in Section 2.1). Originally, Treemaps
represented trees where each leaf node had one weight. However, there are
trees in which leaf nodes have multiple data or time-series data. In order to
draw such trees, many representations derived from Treemaps have been
proposed.
We focus on a representation that embeds data such as bar charts or
area charts in leaf rectangles [2]. When using existing methods, it is
difficult to equalize the scale intervals of charts across the entire Treemap.
As a result, it is confusing and difficult to compare different charts.
We take into consideration the ease of comparison between charts, and
develop a representation in which the widths of leaf rectangles are the
same. In this representation, because the charts have the same width, the
scale intervals of the charts' horizontal axes can be equalized.
Furthermore, to adapt the heights to the weights, it is easy to equalize the
scale intervals of the vertical axes.
2 Background and related work
Many visualization methods for hierarchical structures exist. In this
section, we describe Treemaps and other representations for hierarchical
structures.
2.1 Treemaps
Treemaps are one of the most widely used visualization methods for
hierarchical structures. A Treemap targets rooted trees whose leaf nodes
have some associated weight. Each of these weights is a positive value.
We show an example in Fig. 6.1. Each node is represented as a rectangle,
and child nodes are placed within their parent node. The size of each
rectangle is determined by the weight of the leaf node. A node with a large
weight is represented as a large rectangle. A node with a small weight is
shown as a small rectangle. The size of a branch node is proportional to
the sum of the weights of its child nodes.