Network Visualization in the 'ggplot2' Framework

Network visualization in the 'ggplot2' framework. Network functionality is provided in a single 'ggplot2' layer by calling the geom 'net'. Layouts are calculated using the 'sna' package, example networks are included.

Sam Tyner, Heike Hofmann 2016-12-08 CRAN Status CRAN RStudio mirror downloads Travis-CI Build Status

geomnet is a package built on top of the most recent major ggplot2 release. It provides a ggplot2 geom called geom_net to visualize graphs and networks. It also include the function stat_net to calculate network layouts with the sna package. Finally, the function geom_circle is included to draw circles using ggplot2.

You can install geomnet directly from CRAN install.packages('geomnet') or from Github devtools::install_github("sctyner/geomnet")


ggplot2 Theme Elements

This example shows the theme inheritance properties of the theme elements of ggplot2. Note: this example has not been updated since the release of ggplot2 2.2.0 and as such the content may have changed.

TEnet <- fortify(as.edgedf(theme_elements$edges[,c("parent", "child")]), theme_elements$vertices)
# create a degree variable for use later
TEnet <- TEnet %>%
  group_by(from_id) %>%
  mutate(degree = sqrt(10 * n() + 1))
# plot
ggplot(data = TEnet,
       aes(from_id = from_id, to_id = to_id)) +
  geom_net(layout.alg = "fruchtermanreingold",
    aes(fontsize = degree), directed = TRUE,
    labelon = TRUE, size = 1, labelcolour = 'black',
    ecolour = "grey70", arrowsize = 0.5,
    linewidth = 0.5, repel = TRUE) +
  theme_net() +
  xlim(c(-0.05, 1.05))

Blood Donation Diagram

In this example, we reimagine the traditional blood donation diagram as a directed network. Arrows point to the blood type that receives. This example provides two data frames to geom_net().

ggplot(data = blood$edges, aes(from_id = from, to_id = to)) +
  geom_net(colour = "darkred", layout.alg = "circle", labelon = TRUE, 
           size = 15, directed = TRUE, vjust = 0.5, labelcolour = "grey80",
           arrowsize = 1.5, linewidth = 0.5, arrowgap = 0.05,
           selfloops = TRUE, ecolour = "grey40") + 

Harry Potter Peer Support Network

In this fun example from this website, there is a tie between two students if one provides emotional support to the other at some point in the book. It is a directed network, so in the visualization, the arrow points to the student receiving support.

#>          name1              name2 book
#> 1  Dean Thomas Harry James Potter    1
#> 2  Dean Thomas   Hermione Granger    1
#> 3  Dean Thomas Neville Longbottom    1
#> 4  Dean Thomas     Ronald Weasley    1
#> 5  Dean Thomas    Seamus Finnigan    1
#> 6 Fred Weasley     George Weasley    1
#>                name schoolyear gender      house
#> 1      Adrian Pucey       1989      M  Slytherin
#> 2    Alicia Spinnet       1989      F Gryffindor
#> 3  Angelina Johnson       1989      F Gryffindor
#> 4 Anthony Goldstein       1991      M  Ravenclaw
#> 5     Blaise Zabini       1991      M  Slytherin
#> 6     C. Warrington       1989      M  Slytherin
hp.all <- fortify(as.edgedf(hp.edges), hp.chars, group = "book")
#> Using name1 as the from node column and name2 as the to node column.
#> If this is not correct, rewrite dat so that the first 2 columns are from and to node, respectively.
#> Joining edge and node information by from_id and name respectively.
# only plot the characters with any connections in a given book. 
ggplot(data=hp.all, aes(from_id = from, to_id = to_id)) + 
  geom_net(fiteach=T, directed = T, size = 3, linewidth = .5, 
           ealpha = .5, labelon = T, fontsize = 3, repel = T, 
           labelcolour = "black", arrowsize = .5, singletons = FALSE,
           aes(colour = house, group = house, shape = gender)) + 
  scale_colour_manual(values = c("#941B08","#F1F31C", 
                                 "#071A80", "#154C07")) + 
  facet_wrap(~book, labeller = "label_both") + 
  theme_net() + theme(panel.background = element_rect(colour = 'black'))

Interactivity with plotly!

Now including interactivity with ggplotly()!

# devtools::install_github("ropensci/plotly")
# data step: merge vertices and edges
ftnet <- fortify(as.edgedf(football$edges), football$vertices)
# create data plot
gg <- ggplot(data = ftnet,
       aes(from_id = from_id, to_id = to_id)) +
  geom_net(layout.alg = 'fruchtermanreingold',
    aes(colour = value, group = value,
        linetype = factor(same.conf != 1)),
    linewidth = 0.5,
    size = 5, vjust = -0.75, alpha = 0.3) +
  theme_net() +
  theme(legend.position = "bottom") +
  scale_colour_brewer("Conference", palette = "Paired")  +
  guides(linetype = FALSE)


geomnet 0.2.0

This version of geomnet has been updated to account for changes in ggplot2 version 2.2.0.

Brand New Features

  • Introduction of "adjmat" and "edgedf" classes for newly added fortify methods.
  • fortify methods included for "network", "igraph", "adjmat", and "edgedf" class objects to transform network objects to data frame objects that can be visualized using geom_net().
  • Legend icons are updated to better represent the network data. A micro-network is drawn instead of a single point or line.
  • Added plotly functionality. Hovertext describes edges and provides node labels. Note that arrows are not currently supported in plotly, so directed networks will appear as undirected.
  • New logical parameter singletons added for plotting all nodes in a network even if they have indegree and outdegree of zero. This is most likely to be useful when facetting. The default is TRUE, so all nodes are plotted in all panels. When singletons = FALSE, nodes are only plotted in the panel(s) in which they have at least one incoming or outgoing edge.

Small Usage Changes

  • Plotting labels is slightly different. Use labelon = TRUE to plot labels with the node ids. Use label inside aes() to map a variable value to node label.
  • The selfies argument for drawing self-loops on nodes has been changed to selfloops.

Other file changes

  • Vignette added with several side-by-side comparison examples of geomnet with GGally::ggnet2() and ggnetwork by François Briatte

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.3.1 by Sam Tyner, 7 months ago

Report a bug at

Browse source code at

Authors: Sam Tyner [aut, cre] , Heike Hofmann [aut, ctb] , Nicholas Tierney [ctb]

Documentation:   PDF Manual  

GPL (>= 2) license

Imports sna, network, dplyr, tidyr, readr, plotly

Depends on ggplot2

Suggests ggrepel, knitr, igraph, ggmap, spelling, rmarkdown

See at CRAN