Summarizing Classed Region Maps with a Disk Choreme

TL;DR

This paper presents an algorithmic approach to automatically generate simplified chorematic diagrams from classed region maps, enabling quick and high-quality visual summaries of complex geospatial data.

Contribution

It introduces sampling strategies and algorithms for efficiently creating single-disk chorematic diagrams from large classed region maps, a task previously done mostly manually.

Findings

Effective sampling strategies improve diagram quality

High-quality summaries achieved with moderate point sets

Algorithms run within seconds for large maps

Abstract

Chorematic diagrams are highly reduced schematic maps of geospatial data and processes. They can visually summarize complex situations using only a few simple shapes (choremes) placed upon a simplified base map. Due to the extreme reduction of data in chorematic diagrams, they tend to be produced manually; few automated solutions exist. In this paper we consider the algorithmic problem of summarizing classed region maps, such as choropleth or land use maps, using a chorematic diagram with a single disk choreme. It is infeasible to solve this problem exactly for large maps. Hence, we propose several point sampling strategies and use algorithms for classed point sets to efficiently find the best disk that represents one of the classes. We implemented our algorithm and experimentally compared sampling strategies and densities. The results show that with the right sampling strategy, high-quality results can be obtained already with moderately sized point sets and within seconds of computation time.