Draw a Network / Graph — draw_network • rtemis.draw

Render a network with Sigma.js. Accepts either a square weight / adjacency matrix (e.g. a correlation matrix) or an edge-list data frame, dispatching on the type of x. Communities are detected with Louvain and laid out with ForceAtlas2 by default.

Usage

draw_network(
  x,
  nodes = NULL,
  directed = FALSE,
  threshold = NULL,
  self_loops = FALSE,
  layout = "force",
  node_size = 10,
  edge_scale = 3,
  node_opacity = 0.95,
  edge_opacity = 0.4,
  show_labels = TRUE,
  scale_by_degree = TRUE,
  color_by_group = FALSE,
  resolution = 1,
  blend_edges = FALSE,
  palette = palette_colors(rtemis_colors),
  node_color = rtemis_colors[["teal"]],
  positive_color = rtemis_colors[["teal"]],
  negative_color = "#ff9e1f",
  title = NULL,
  theme = NULL,
  width = NULL,
  height = NULL,
  element_id = NULL,
  filename = NULL
)

Arguments

x: Numeric matrix or data.frame: A square weight / adjacency matrix, or an edge-list data frame with source / target columns (and optional weight, sign). For a matrix, dimnames supply node ids and [i, j] is the edge weight (zero / NA entries produce no edge); the absolute value drives thickness and the sign drives color. For an edge list, source and target name the endpoints (the first two columns if those names are absent), with optional weight and sign columns.
nodes: Optional Data frame: Node attributes (id/name, label, value, group). Used only with edge-list input; ignored for matrix input.
directed: Logical: Whether the graph is directed.
threshold: Optional Numeric [0, Inf): For matrix input, drop edges whose absolute weight is below this value.
self_loops: Logical: For matrix input, whether to keep diagonal self-edges.
layout: Character {"force", "circular", "circlepack", "random"}: Layout algorithm. "force" is ForceAtlas2.
node_size: Numeric [0, Inf): Base node radius in screen pixels.
edge_scale: Numeric [0, Inf): Multiplier mapping normalized edge weight to stroke width.
node_opacity: Numeric [0, 1]: Node fill opacity.
edge_opacity: Numeric [0, 1]: Edge stroke opacity.
show_labels: Logical: Whether to render node labels.
scale_by_degree: Logical: Scale each node's radius by its value (weighted degree); when FALSE all nodes use the base size.
color_by_group: Logical: Color nodes by detected community (Louvain) instead of a single hue.
resolution: Numeric [0, Inf): Louvain resolution; higher yields more, smaller communities.
blend_edges: Logical: Color each edge as the blend of its two endpoint node colors instead of by sign.
palette: Character: Categorical colors for communities.
node_color: Character: Single node color used when color_by_group is FALSE.
positive_color: Character: Edge color for positive-sign edges.
negative_color: Character: Edge color for negative-sign edges.
title: Optional Character: Title shown above the network.
theme: Optional Theme, list, or NA: Theme override. NULL enables light/dark auto-detection (matching draw()).
width: Optional Character or Numeric: Widget width.
height: Optional Character or Numeric: Widget height.
element_id: Optional Character: Explicit element ID.
filename: Optional Character: Static export of network widgets is not yet supported. Supplying a path raises an error.

Value

htmlwidget.

Examples

# Correlation matrix
draw_network(cor(mtcars), threshold = 0.5)


# Edge list
edges <- data.frame(
  source = c("a", "a", "b"),
  target = c("b", "c", "c"),
  weight = c(1, 2, 0.5)
)
draw_network(edges)