treemaps

A treemap is a space-filling visualization for hierarchical data that represents each node as a rectangle. The area of each rectangle corresponds to a quantitative attribute, such as size or value, while the rectangles are nested to reflect the hierarchy. Colors can encode categories or another metric.

Treemaps are created by partitioning a region into rectangles corresponding to the root's children, recursively applying the same idea to their descendants. Several layout algorithms exist, including slice-and-dice, which alternates horizontal and vertical cuts; and squarified treemaps, which strive for near-square rectangles to improve readability.

The concept was popularized by Ben Shneiderman in 1991. A major advance was the squarified treemap proposed by Bruls, Huizing and van Wijk in 2000, which reduces extreme aspect ratios and makes it easier to compare values within a level.

Treemaps are used to explore large hierarchies such as file systems, software structures, financial portfolios, or taxonomies. They support interactive exploration through zooming or breadcrumb navigation and can include labels, color scales, and tooltips.

Benefits include compact visualization of large hierarchies and immediate visual impression of relative magnitudes. Limitations include small or obscured rectangles for deep trees, potential misinterpretation of area, and difficulty reading precise values. Alternatives such as icicle charts or sunbursts address some issues.

Variants and extensions incorporate interactivity, animation, or focus+context techniques to mitigate limitations, while combining treemaps with additional encodings. Common software tools and libraries provide treemap implementations for data analysis, business intelligence, and educational visualizations.