Sankey / sunburst input from treatment-line data

Turn one-row-per-subject-per-line data (e.g. lines of therapy) into the inputs of tfl_plot_sankey() and tfl_plot_sunburst().

Usage

tfl_sankey_data(
  data,
  id = "USUBJID",
  stage = "LINE",
  category = "TRT",
  by = NULL,
  levels = NULL,
  stage_prefix = "L"
)

tfl_sunburst_data(
  data,
  id = "USUBJID",
  stage = "LINE",
  category = "TRT",
  levels = NULL,
  stage_prefix = "L"
)

Arguments

data

Data frame with one row per subject and stage (line).

id

Subject variable.

stage

Stage variable (e.g. line number); stages are ordered by their values (numeric order when numeric).

category

Category drawn as nodes / arcs (e.g. treatment class).

by

Optional subgroup variable: nodes and links are built for all subjects ("ALL") and for each value of by, marked in column grp.

levels

Optional order of the categories (default: order of first appearance, stage by stage).

stage_prefix

Prefix of the node labels ("L" gives "L1: Chemo").

Value

tfl_sankey_data(): a list with nodes (id, stage, category, label, n, and grp when by is given) and links (source, target, value, grp). tfl_sunburst_data(): a data frame with one column per stage (L1, L2, ...) and the subject count n.

Examples

lot <- data.frame(
  USUBJID = c("1", "1", "2", "3", "3", "3"),
  LINE    = c(1, 2, 1, 1, 2, 3),
  TRT     = c("Chemo", "IO", "Chemo", "IO", "Chemo", "Targeted")
)
sk <- tfl_sankey_data(lot, stage = "LINE", category = "TRT")
tfl_plot_sankey(sk$nodes, sk$links, node_label = "label", node_value = "n",
            node_treatment = "category")

tfl_plot_sunburst(tfl_sunburst_data(lot, stage = "LINE", category = "TRT"))