Description
The package implements several new layout algorithms to visualize networks that are not part of ‘igraph’. Most are based on the concept of stress majorization by Gansner et al. (2004) doi:10.1007/978-3-540-31843-9_25. Some more specific algorithms allow the user to emphasize hidden group structures in networks or focus on specific nodes.
Keywords
- Network Analysis
- Network Visualization
- Network layouts
Use Cases
Network visualization offers social scientists a powerful tool for analyzing relationships and interactions within digital traces. For instance, researchers studying online communities can use network visualization to map interactions on social media platforms, such as X or Reddit. By visualizing user interactions (like replies, mentions, or shared links), researchers can uncover patterns in information flow, identify influential users, and explore the formation of communities or echo chambers. Network visualization can reveal clusters of users who frequently engage with one another, suggesting tightly-knit subgroups with shared interests or beliefs. It also helps identify key influencers within these networks, who may play a critical role in spreading information or shaping public opinion. This analysis is particularly useful for understanding phenomena like misinformation spread, public discourse on sensitive topics, or the social dynamics of online activism, offering insights into how ideas and behaviors propagate through digital spaces.
Input Data
graphlayouts accepts igraph network objects as input and includes many datasets that can be used to test the layout algorithms.
Output Data
The output of graphlayouts is a layout object created with create_layout() from ggraph.
This object is essentially a data frame that contains the x–y coordinates of each node, along with any node attributes from the input graph. It can be directly passed into plotting functions in ggraph.
Environment Setup
With R installed:
install.packages("graphlayouts")Installing package into '/srv/rlibs'
(as 'lib' is unspecified)How to Use
After installing the package and its dependencies, you can start visualizing networks with graphlayouts and the ggraph plotting framework.
Create or load a graph
Graphs can be created manually or loaded from existing datasets. For example:
library(igraph)
Attaching package: 'igraph'The following objects are masked from 'package:stats':
decompose, spectrumThe following object is masked from 'package:base':
uniong <- make_ring(10) # simple ring graph with 10 nodesChoose a layout algorithm
The core contribution of graphlayouts is a collection of layout algorithms not included in base igraph. Layouts are deterministic (they produce the same result each time) and often based on stress majorization, which aims to preserve graph-theoretic distances in 2D space.
Use create_layout() from ggraph with one of the following layout options provided by graphlayouts: - "stress" – general-purpose layout that minimizes stress in node placement.
- "focus" – emphasizes the position of one or more focal nodes.
- "backbone" – highlights the backbone structure of a network while de-emphasizing weaker connections.
- "centrality" – arranges nodes according to a chosen centrality measure (e.g., degree or betweenness).
library(graphlayouts)
library(ggraph)Loading required package: ggplot2lay <- create_layout(g, layout = "stress")Draw the network with ggraph
With a layout prepared, you can visualize the graph using ggraph layers.
- Nodes: use
geom_node_point()for basic dots,geom_node_text()for labels, or scale aesthetics (size, color, fill) by node attributes.
- Edges: use
geom_edge_link()for straight lines,geom_edge_arc()for curved arcs, or directional edges with arrows.
ggraph(lay) +
geom_edge_link(alpha = 0.5) +
geom_node_point(size = 5, color = "steelblue") +
theme_graph()
Customize appearance
You can customize almost every aspect of the plot: - Node size or color can map to graph metrics:
V(g)$deg <- degree(g)
lay <- create_layout(g, layout = "stress")
ggraph(lay) +
geom_edge_link(alpha = 0.3) +
geom_node_point(aes(size = deg), color = "tomato") +
theme_graph()
- Edge thickness or color can map to weights or categories.
- Labels can be added with
geom_node_text()orgeom_node_label().
- Themes and palettes: use
theme_graph()for a clean default or apply any ggplot2 scales (e.g.,scale_fill_viridis_c()).
Explore alternatives
graphlayouts is particularly useful when different research questions call for different visual emphases: - Use "focus" layouts to visualize networks from the perspective of a key actor.
- Use "backbone" layouts to highlight community structures.
- Compare multiple layouts to see how structural features (clusters, bridges, hubs) appear under different perspectives.
You can find a complete tutorial here
Technical Details
See the official CRAN page for further information about technical details.
Contact Details
Maintainer: David Schoch david@schochastics.net
Issue Tracker: https://github.com/schochastics/graphlayouts/issues