To draw a tree

Abstract

The quintessential goal of information visualization is depicting abstractions and relations for non-spatial data. A hierarchy is a particularly expressive abstraction that can be applied to a broad range of domains: the genealogical lineages of human descent, the functional decomposition of complex mechanical objects, the classification of knowledge, the evolutionary relationships between species. All of these hierarchical relationships are representable through the abstraction of a recursively defined tree. For this reason, trees occupy a place along with arrays, lists and graphs as one of the most important data structures in computer science. Considering the simple problem of how to effectively draw a tree uncovers many issues fundamental to information visualization. Different drawing styles emphasize different properties of trees, often in subtle ways. I will discuss how people think about trees, and thus what kinds of relationships a tree drawing can usefully convey. My discussion will include a review of many methods for drawing trees, including both historical examples from the sciences and techniques recently developed by researchers in information visualization.