Space Partitioning Schemes and Algorithms for Generating Regular and Spiral Treemaps

TL;DR

This paper introduces a new optimization model and five algorithms for generating high-quality treemaps, including divide and conquer and spiral methods, improving visual appeal and computational efficiency.

Contribution

It presents a novel optimization framework and five algorithms, expanding the capabilities and efficiency of treemap generation methods.

Findings

Optimization model produces superior treemaps for benchmarking.

Algorithms improve aspect ratio and stability of treemaps.

Four algorithms are computationally efficient with quasilinear time.

Abstract

Treemaps have been widely applied to the visualization of hierarchical data. A treemap takes a weighted tree and visualizes its leaves in a nested planar geometric shape, with sub-regions partitioned such that each sub-region has an area proportional to the weight of its associated leaf nodes. Efficiently generating visually appealing treemaps that also satisfy other quality criteria is an interesting problem that has been tackled from many directions. We present an optimization model and five new algorithms for this problem, including two divide and conquer approaches and three spiral treemap algorithms. Our optimization model is able to generate superior treemaps that could serve as a benchmark for comparing the quality of more computationally efficient algorithms. Our divide and conquer and spiral algorithms either improve the performance of their existing counterparts with respect to aspect ratio and stability or perform competitively. Our spiral algorithms also expand their applicability to a wider range of input scenarios. Four of these algorithms are computationally efficient as well with quasilinear running times and the last algorithm achieves a cubic running time. A full version of this paper with all appendices, data, and source codes is available at \anonymizeOSF{\OSFSupplementText}.