make_sankey: Create an object that describes a sankey plot

In sankey: Illustrate the Flow of Information or Material

Description

Create an object that describes a sankey plot

Usage

make_sankey(nodes = NULL, edges, y = c("optimal", "simple"),
  break_edges = FALSE, gravity = c("center", "top", "bottom"))

Arguments

nodes	A data frame of nodes on the plot, and possibly their visual style. The first column must be the ids of the nodes. If this argument is NULL, then the ids of the nodes are determined from edges.
edges	A data frame of the edges. The first two columns must be node ids, and they define the edges. The rest of the columns contain the visual style of the edges.
y	How to calculate vertical coordinates of nodes, if they are not given in the input. optimal tries to minimize edge crossings, simple simply packs nodes in the order they are given, from bottom to top.
break_edges	Whether to plot each edge as two segments, or a single one. Sometimes two segment plots look better.
gravity	Whether to push the nodes to the top, to the bottom or to the center, within a column.

Details

The node and edges data frames may contain columns that specify how the plot is created. All parameters have reasonable default values.

Current list of graphical parameters for nodes:

• col Node color.

• size Node size.

• x Horizontal coordinates of the center of the node.

• y Vertical coordinates of the center of the node.

• shape Shape of the node. Possible values: rectangle, point, invisible.

• lty Lite type, see par.

• srt How to rotate the label, see par.

• textcol Label color.

• label Label text. Defaults to node name.

• adjx Horizontal adjustment of the label. See adj in the par manual.

• adjy Vertical adjustment of the label. See adj in the par manual.

• boxw Width of the node boxes.

• cex Label size multiplication factor.

• top Vertical coordinate of the top of the node.

• center Vertical coordinate of the center of the node.

• bottom Vertical coordinate of the bottom of the node.

• pos Position of the text label, see par.

• textx Horizontal position of the text label.

• texty Vertical position of the text label.

Current list of graphical parameters for edges:

• colorstyle Whether the to use a solid color (col), or gradient to plot the edges. The color of a gradient edges is between the colors of the nodes.

• curvestyle Edge style, sin for sinusoid curves, line for straight lines.

• col Edge color, for edges with solid colors.

• weight Edge weight. Determines the width of the edges.

Value

A sankey object that can be plotted via the sankey function.x

Examples

## Function calls in the pkgsnap package:
edges <- read.table(stringsAsFactors = FALSE, textConnection(
"                get_deps          get_description
                 get_deps               parse_deps
                 get_deps                     %||%
                 get_deps            drop_internal
          get_description        pkg_from_filename
               parse_deps                 str_trim
                cran_file             get_pkg_type
                cran_file          r_minor_version
            download_urls split_pkg_names_versions
            download_urls                cran_file
             pkg_download               dir_exists
             pkg_download            download_urls
             pkg_download        filename_from_url
             pkg_download             try_download
                  restore             pkg_download
                  restore        drop_missing_deps
                  restore            install_order
                  restore                 get_deps
 split_pkg_names_versions               data_frame
"))
pkgsnap_sankey <- make_sankey(edges = edges)
sankey(pkgsnap_sankey)

## Some customization
nodes <- data.frame(
  stringsAsFactors = FALSE,
  id = c("snap", sort(unique(c(edges[,1], edges[,2]))))
)
nodes$col <- ifelse(nodes$id %in% c("snap", "restore"), "orange", "#2ca25f")
edges$colorstyle <- "gradient"

sankey(make_sankey(nodes, edges))

Example output

sankey documentation built on May 2, 2019, 11:07 a.m.