The partition layout is a way to show hierarchical data in the same way as layout_tbl_graph_treemap(). Instead of subdividing the parent area the partition layout shows the division of a nodes children next to the area of the node itself. As such the node positions will be very reminiscent of a reingold-tilford tree layout but by plotting nodes as areas it better communicate the total weight of a node by summing up all its children. Often partition layouts are called icicle plots or sunburst diagrams (in case a radial transform is applied).

layout_tbl_graph_partition(
  graph,
  weight = NULL,
  circular = FALSE,
  height = NULL,
  sort.by = NULL,
  direction = "out",
  offset = pi/2,
  const.area = TRUE
)

Arguments

graph

An tbl_graph object

weight

An optional node variable to use as weight. Will only affect the weight of leaf nodes as the weight of non-leaf nodes are derived from their children.

circular

Logical. Should the layout be transformed to a circular representation. If TRUE the resulting layout will be a sunburst diagram.

height

An optional node variable to use as height. If NULL all nodes will be given a height of 1.

sort.by

The name of a node variable to sort the nodes by.

direction

The direction of the tree in the graph. 'out' (default) means that parents point towards their children, while 'in' means that children point towards their parent.

offset

If circular = TRUE, where should it begin. Defaults to pi/2 which is equivalent to 12 o'clock.

const.area

Logical. Should 'height' be scaled for area proportionality when using circular = TRUE. Defaults to TRUE.

Value

If circular = FALSE A data.frame with the columns x, y, width, height, leaf, depth, circular as well as any information stored as node variables in the tbl_graph object. If circular = TRUE A data.frame with the columns x, y, r0, r, start, end, leaf, depth, circular as well as any information stored as node variables in the tbl_graph object.

Note

partition is a layout intended for trees, that is, graphs where nodes only have one parent and zero or more children. If the provided graph does not fit this format an attempt to convert it to such a format will be made.

References

Kruskal, J. B., Landwehr, J. M. (1983). Icicle Plots: Better Displays for Hierarchical Clustering. American Statistician Vol 37(2), 162-168. https://doi.org/10.2307/2685881