PieGlyph: An R package for creating axis invariant pie-glyphs for 2d plots

Abstract

Effective visualisation of multidimensional data is crucial for generating insights. Glyph-based visualisations, which encode data dimensions onto multiple visual channels such as colour, shape, and size, provide an effective means of representing complex datasets. Pie-chart glyphs (pie-glyphs) are one such approach, where multiple data attributes are mapped to slices within a pie chart. This paper introduces the PieGlyph R package, which enables users to overlay any 2D plot with axis-invariant pie-glyphs, offering a compact and intuitive representation of multidimensional data. Unlike existing R packages such as scatterpie or ggforce, PieGlyph generates pie-glyphs independently of the plot axes by employing a nested coordinate system, ensuring they remain circular regardless of changes to the underlying coordinate system. This enhances interpretability, particularly in when visualising spatial data, as users can select the most appropriate map projection without distorting the glyphs' shape. Pie-glyphs are also particularly well-suited for visualising compositional data, where there is a natural sum-to-one constraint on the data attributes. PieGlyph is developed under the Grammar of Graphics paradigm using the ggplot2 framework and supports the generation of interactive pie-glyphs through the ggiraph package. Designed to integrate seamlessly with all features and extensions offered by ggplot2 and ggiraph, PieGlyph provides users with full flexibility in customising every aspect of the visualisation. This paper outlines the conceptual framework of PieGlyph, compares it with existing alternatives, and demonstrates its applications through example visualisations.