scale_edge_widths: Scale Edge Widths Based on Weights

In cograph: Analysis and Visualization of Complex Networks

Scale Edge Widths Based on Weights

Description

Unified edge width scaling function that supports multiple scaling modes, two-tier cutoff system (like qgraph), and output range specification.

Usage

scale_edge_widths(
  weights,
  esize = NULL,
  n_nodes = NULL,
  directed = FALSE,
  mode = "linear",
  maximum = NULL,
  minimum = 0,
  cut = NULL,
  range = c(0.5, 4)
)

Arguments

weights	Numeric vector of edge weights.
esize	Base edge size. NULL uses adaptive sizing based on n_nodes.
n_nodes	Number of nodes (for adaptive esize calculation).
directed	Whether network is directed (affects adaptive esize).
mode	Scaling mode: "linear", "log", "sqrt", or "rank".
maximum	Max weight for normalization. NULL for auto-detect.
minimum	Min weight threshold. Edges below this get minimum width.
cut	Two-tier cutoff threshold. NULL = auto (75th percentile), 0 = disabled (continuous scaling), positive number = manual threshold.
range	Output width range as c(min_width, max_width).

Details

Scaling Modes

• linear (default): Direct proportional scaling, matches qgraph behavior.

• log: Logarithmic scaling for wide weight ranges. Uses log1p for stability.

• sqrt: Square root scaling for moderate compression.

• rank: Rank-based scaling for equal visual spacing regardless of weight distribution.

Two-Tier System (cut parameter)

When cut > 0, edges are divided into two tiers:

• Below cut: Minimal width variation (20% of range)

• Above cut: Full width scaling (80% of range)

This matches qgraph's behavior where weak edges are visually de-emphasized.

Value

Numeric vector of scaled edge widths.

cograph documentation built on April 1, 2026, 1:07 a.m.