cograph

cograph is a modern R package for the analysis and visualization of complex networks, designed for simplicity, tidy outputs, comprehensive statistics and up-to-date network science. cograph accepts matrices, edge lists, and igraph, statnet, qgraph and tna objects without conversion, and offers a wide array of tools for plotting, wrangling, centrality, community detection, motif, robustness, multilayer and higher-order analysis.

Installation

# Release version from CRAN
install.packages("cograph")

# Development version from GitHub
# install.packages("remotes")
# remotes::install_github("sonsoleslp/cograph")
install.packages("cograph", repos = c('https://sonsoleslp.r-universe.dev', 'https://cloud.r-project.org'))

Quick start

The examples use regulation_net, a synthetic weighted transition network among ten learning states included in the package. splot() plots it in one call, and tna_styling = TRUE applies the visual conventions of transition networks.

library(cograph)
splot(regulation_net, tna_styling = TRUE)

centrality() returns any combination of measures as a tidy data frame, from the classical measures to recent ones such as randomized shortest-path betweenness and Trust-PageRank.

centrality(regulation_net,
           measures = c("strength", "betweenness", "pagerank",
                        "rsp_betweenness", "trust_pagerank"),
           sort_by = "pagerank", digits = 3)
#>          node strength_all betweenness pagerank rsp_betweenness trust_pagerank
#> 1     Monitor         1.87        18.0    0.184         132.027          0.147
#> 2      Create         1.64        13.0    0.138         102.913          0.117
#> 3     Reflect         1.39        10.0    0.125          79.405          0.084
#> 4       Adapt         1.77        15.0    0.124          91.887          0.112
#> 5     Explore         1.39         5.0    0.118          79.715          0.086
#> 6       Share         1.95         9.0    0.095          69.907          0.092
#> 7    Evaluate         1.71         3.0    0.074          51.387          0.092
#> 8     Discuss         1.53         0.5    0.068          43.823          0.088
#> 9  Synthesize         0.77         6.5    0.038          23.304          0.067
#> 10       Plan         1.90        15.5    0.036          22.235          0.114

plot_mcml() shows a network whose nodes belong to clusters as a two-layer hierarchy, with the node-level network below and the cluster-level network above.

clusters <- list(Cognitive  = c("Explore", "Plan", "Monitor", "Adapt", "Reflect"),
                 Social     = c("Discuss", "Synthesize", "Share"),
                 Evaluative = c("Evaluate", "Create"))
plot_mcml(regulation_net, clusters)

plot_simplicial() visualizes higher-order pathways over the network, with each pathway joining the states that lead to a target state.

plot_simplicial(regulation_net,
                c("Explore Plan -> Monitor", "Monitor Adapt -> Reflect",
                  "Discuss Synthesize -> Evaluate", "Create Share -> Explore"))

What cograph covers

• Visualization. splot() plots any supported input with specialized styling for transition and psychological networks, alongside a wide array of specialized plots from alluvial flows and chord diagrams to bootstrap forest plots and temporal prisms.
• Wrangling. cograph offers a family of wrangling verbs for selecting, filtering, thresholding, transforming and editing networks, each returning a network so that the verbs chain with the native pipe.
• Centrality. centrality() returns a large collection of node centrality measures across all major families as a tidy data frame, tested against igraph, sna, centiserve, NetworkX and other implementations where they exist.
• Network statistics. network_summary() returns density, diameter, centralization, reciprocity, transitivity and many further statistics in one data frame.
• Communities. communities() runs a range of detection algorithms through one call, with consensus, comparison and significance testing of partitions.
• Motifs. motifs() and subgraphs() count the triads of the MAN classification, test their frequencies and identify the nodes that form each pattern.
• Robustness. robustness() and vulnerability() simulate targeted and random attacks and measure each node’s contribution to the efficiency of the network.
• Clusters, layers and higher-order structure. cograph offers hierarchical plots for multi-cluster networks, supra-adjacency tools for multilayer networks, and visualization of higher-order pathways estimated with Nestimate.

Documentation

Tutorials

• Network visualization with cograph: a complete guide to plotting with splot().
• Communities and higher-order networks: detecting communities and visualizing them over the network.
• Network estimation with Nestimate and cograph: from sequence data to bootstrapped, compared and clustered networks.
• Multi-cluster multi-level visualization: hierarchical plots of clustered networks with plot_mcml().
• Higher-order network analysis with simplicial complexes: from transition networks to topological analysis.

Articles

• Introduction to cograph: an overview of the package.
• Why cograph?: the design of the package.
• Centrality catalogue: every centrality measure with its definition and interpretation.
• cograph and the Centrality Zoo: cograph’s measures compared with the Zoo and with other packages.
• Plotting TNA models: a gallery of TNA plots.
• Advanced MCML examples: further multi-cluster figures.
• Bootstrap forest plots: confidence intervals of bootstrapped edges.
• Migrating from qgraph to splot: qgraph arguments and their cograph equivalents.

Citation and license

Please cite cograph with citation("cograph"). cograph is released under the MIT license.