Plot a temporal hypergraph — plot.net_temporal

Plots the snapshot that hg_snapshot() takes at at as a hypergraph, with plot.net_hg(): the hyperedges active then, as hulls by default or as the incidence matrix with type = "incidence". The snapshot keeps the hyperedge attributes and the data's names for nodes and hyperedges, so color_by can name an attribute and the legends use the data's words.

Usage

# S3 method for class 'net_temporal_hypergraph'
plot(
  x,
  at = NULL,
  mode = c("active", "cumulative"),
  type = c("hulls", "incidence", "storyline"),
  top = 8L,
  start = NULL,
  end = NULL,
  method = NULL,
  ...
)

Arguments

x: A temporal_hypergraph().
at: Snapshot time; defaults to the end of observation. Not used by the storyline.
mode: Snapshot mode passed to hg_snapshot(). Not used by the storyline.
type: "hulls" (default) or "incidence" plot the snapshot at at (see plot.net_hg()); "storyline" plots the whole history.
top: For type = "storyline": the number of nodes with the most hyperedges to draw as lines (default 8, one Okabe-Ito colour each; ties broken by name), or NULL for every node.
start, end: For type = "storyline": draw only the hyperedges that begin in this period, as dates for a calendar hypergraph or numbers on its clock. NULL (default) leaves the period open.
method: Deprecated. The snapshot used to be projected to a pairwise network; plot(pairwise_network(hg_snapshot(x, at), type = method)) plots that network.
...: For "hulls" and "incidence", arguments passed to plot.net_hg(), such as color_by or labels. For "storyline", edge_labels (TRUE writes the hyperedge and its start under each column), point_size (size of the points, default 2.5) and spacing: "even" (default) puts neighbouring lines one row apart, "strength" brings two neighbouring lines closer the more plotted hyperedges their nodes share: a full row apart for none, 0.35 of a row for the largest number shared by any pair in the plot, and linearly in between, so lines that often meet run together. Spacing never changes the order of the lines. width_by = "degree" draws each line with a width that grows linearly with its node's number of hyperedges in the period, the count top ranks by, and adds a width legend; NULL (default) draws every line alike.

Value

A ggplot object.

Details

type = "storyline" plots the whole history instead (Tanahashi and Ma 2012). Each of the top nodes with the most hyperedges is a line, from its first hyperedge to its last, and each hyperedge is a column, in order of its start, where a grey bar gathers the lines of its members among the plotted nodes. The columns are spaced by order, not by elapsed time, and are labelled with the hyperedge and its start. Lines are ordered by the barycentre rule of Sugiyama, Tagawa and Toda (1981): along the columns the members of each hyperedge move to the median of their current rows, and the ordering is refined over repeated sweeps, keeping the one with the fewest line crossings. Each line takes an Okabe-Ito colour and its points a shape, and the legend names the nodes in decreasing order of their number of hyperedges. Eight colours and nine shapes cycle, so up to 72 lines differ in their pair of colour and shape.

Conditions

hypergraphs_bad_input when the snapshot has no active nodes, when no hyperedge begins between start and end, for an invalid top, for at, mode or an argument of plot.net_hg() with type = "storyline", for top, start or end with another type, and as plot.net_hg() raises it; hypergraphs_deprecated (a warning) for method.

References

Tanahashi, Y., & Ma, K.-L. (2012). Design considerations for optimizing storyline visualizations. IEEE Transactions on Visualization and Computer Graphics, 18(12), 2679-2688. doi:10.1109/TVCG.2012.212

Sugiyama, K., Tagawa, S., & Toda, M. (1981). Methods for visual understanding of hierarchical system structures. IEEE Transactions on Systems, Man, and Cybernetics, 11(2), 109-125. doi:10.1109/TSMC.1981.4308636

Examples

seats <- data.frame(
  case = rep(c("A", "B", "C"), each = 3),
  arbitrator = c("p1", "a1", "a2", "p2", "a1", "a3", "p1", "a4", "a5"),
  constituted = rep(c(1, 2, 4), each = 3),
  concluded = rep(c(4, 3, 6), each = 3)
)
thg <- temporal_hypergraph(seats, node = "arbitrator", hyperedge = "case",
                           start = "constituted", end = "concluded")
plot(thg, at = 2.5)

plot(thg, at = 2.5, type = "incidence", edge_labels = TRUE)

plot(thg, type = "storyline")