Chapter 6 edges

6.1 Edge attributes

On the basis of data from Lemurs (!) we’ll see which attributes we can vary from edges.

Data from this wonderful website.

lemur_graphml <- read.graph("https://raw.githubusercontent.com/bansallab/asnr/master/Networks/Mammalia/lemur_sexual_weighted/lemur_pereira_sexual_network_1_attribute.graphml", format = "graphml")

lemur <- as_tbl_graph(lemur_graphml)

lemur
## # A tbl_graph: 9 nodes and 7 edges
## #
## # An unrooted forest with 2 trees
## #
## # Node Data: 9 × 3 (active)
##   `residence status` SEX    id   
##   <chr>              <chr>  <chr>
## 1 "immigrant"        MALE   A    
## 2 "resident"         FEMALE TR   
## 3 ""                 MALE   G    
## 4 "resident"         FEMALE CN   
## 5 "resident"         FEMALE LY   
## 6 "immigrant"        MALE   N    
## # … with 3 more rows
## #
## # Edge Data: 7 × 3
##    from    to weight
##   <int> <int>  <dbl>
## 1     1     2      1
## 2     3     4      1
## 3     3     5      1
## # … with 4 more rows

The standard graph would look something like this:

ggraph(lemur, layout = "kk") +
  geom_edge_link() +
  geom_node_point(size = 4) +
  theme_graph()

6.1.1 edge colour

An important distinction is that between “setting” and “mapping” variables. Let’s use the colour of edges to see the difference. We could simply ‘set’ the colour of the lines in the following way:

ggraph(lemur, layout = "kk") +
  geom_edge_link(colour = "purple") +
  geom_node_point(size = 4) +
  theme_graph()

A slightly more interesting thing would be to map a variable to the colour of the lines. For example:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(colour = factor(weight))) +
  geom_node_point(size = 4) +
  theme_graph()

Those colours are not so nice, let’s try to choose our own colours:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(colour = factor(weight))) +
  geom_node_point(size = 4) +
  theme_graph() +
  scale_edge_colour_brewer(palette = "Set1")

6.1.2 edge width

We can do the same tricks for line width as we did for colour:

Setting the width:

ggraph(lemur, layout = "kk") +
  geom_edge_link(width = 4, colour = "grey") +
  geom_node_point(size = 4) +
  theme_graph()

Mapping a variable to width:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(width = factor(weight)), colour = "grey") +
  geom_node_point(size = 4) +
  theme_graph()

6.1.3 linetype

Setting:

ggraph(lemur, layout = "kk") +
  geom_edge_link(linetype = "dashed") +
  geom_node_point(size = 4) +
  theme_graph()

Mapping:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(linetype = factor(weight))) +
  geom_node_point(size = 4) +
  theme_graph()

6.1.4 seethroughness

Setting:

ggraph(lemur, layout = "kk") +
  geom_edge_link(alpha = 0.5) +
  geom_node_point(size = 4) +
  theme_graph()

Mapping:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(alpha = factor(weight))) +
  geom_node_point(size = 4) +
  theme_graph()

As we’ve seen, alpha is useful when you have many, many edges:

dolphins_graphml <- igraph::read.graph("https://raw.githubusercontent.com/bansallab/asnr/master/Networks/Mammalia/dolphin_association_weighted/weighted_FORAGE_dolphin_florida.graphml", format = "graphml")

dolphins <- as_tbl_graph(dolphins_graphml)

ggraph(dolphins, layout = "kk") +
  geom_edge_link(alpha = 0.1) +
  geom_node_point() +
  theme_graph()

Another cool trick of alpha:

ggraph(lemur, layout = "kk") +
  geom_edge_link(aes(alpha = ..index..), show.legend = FALSE) +
  geom_node_point(size = 4) +
  theme_graph()

6.1.5 beginning and the end

6.1.5.1 Arrows

We’ll see some arrows:

ggraph(lemur, layout = "kk") +
  geom_edge_link(
    colour = "purple",
    arrow = arrow(length = unit(4, 'mm')) 
  ) +
  geom_node_point(size = 4) +
  theme_graph()

ggraph(lemur, layout = "kk") +
  geom_edge_link(
    colour = "grey",
    arrow = arrow(angle = 45, length = unit(7, 'mm'), type = "closed") 
  ) +
  geom_node_point(size = 4) +
  theme_graph()

It’s ok, but we would like to avoid overlap between the edges and nodes, which we can achieve by start_cap and end_cap:

ggraph(lemur, layout = "kk") +
  geom_edge_link(
    colour = "grey",
    arrow = arrow(angle = 45, length = unit(4, 'mm'), type = "closed"),
    end_cap = circle(3, 'mm')  
  ) +
  geom_node_point(size = 4) +
  theme_graph()

Let’s do the same for the start

ggraph(lemur, layout = "kk") +
  geom_edge_link(
    colour = "grey",
    arrow = arrow(angle = 45, length = unit(4, 'mm'), type = "closed"),
    start_cap = circle(6, 'mm'),
    end_cap = circle(3, 'mm')  
  ) +
  geom_node_point(size = 4) +
  theme_graph()

6.2 Different types of edges

Within the package ggraph, we can use the following functions to create different kinds of edges. Some edge types are only useful or particular layouts. :

6.2.2 geom_edge_arc

Bendy bendy, so overlapping lines are visible.

ggraph(lemur, layout = "kk") +
  geom_edge_arc() +
  geom_node_point(size = 4) +
  theme_graph()

In the above case they’re not so useful, let’s find a better example:

ggraph(lemur, layout = "linear") +
  geom_edge_arc(colour = "grey") +
  geom_node_point(size = 4) +
  theme_graph()

ggraph(lemur, layout = "linear", circular = TRUE) +
  geom_edge_arc(colour = "grey") +
  geom_node_point(size = 4) +
  theme_graph() +
  coord_fixed()

6.2.3 geom_edge_bend

Particularly useful for hierarchical plots.

ggraph(lemur, layout = "tree") +
  geom_edge_bend() +
  geom_node_point(size = 4) +
  theme_graph()

6.2.4 geom_edge_diagonal

Also useful for hierarchical plots

ggraph(lemur, layout = "tree") +
  geom_edge_diagonal() +
  geom_node_point(size = 4) +
  theme_graph()

6.2.5 geom_edge_fan

The fan is useful when you have directed arrows between two nodes. Let’s revise our data on lemur by drawing additional directional ties. We’ll draw a tie from 2 to 1, rom 8 to 7, and from 1 to 1:

lemur2 <- lemur %>% 
  convert(to_directed) %>%
  bind_edges(data.frame(from = c(2, 8, 1), to = c(1, 7, 1), weight = c(1, 1, 1)))  

lemur2 %>% activate(edges) %>% arrange(from)
## # A tbl_graph: 9 nodes and 10 edges
## #
## # A directed multigraph with 2 components
## #
## # Edge Data: 10 × 4 (active)
##    from    to weight .tidygraph_edge_index
##   <int> <int>  <dbl>                 <int>
## 1     1     2      1                     1
## 2     1     1      1                    NA
## 3     2     1      1                    NA
## 4     3     4      1                     2
## 5     3     5      1                     3
## 6     4     8      1                     4
## # … with 4 more rows
## #
## # Node Data: 9 × 4
##   `residence status` SEX    id    .tidygraph_node_index
##   <chr>              <chr>  <chr>                 <int>
## 1 "immigrant"        MALE   A                         1
## 2 "resident"         FEMALE TR                        2
## 3 ""                 MALE   G                         3
## # … with 6 more rows

With geom_edge_link we don’t see the directed ties!

ggraph(lemur2, layout = "kk") +
  geom_edge_link() +
  geom_node_point(size = 4) +
  theme_graph()

geom_edge_fan to save the day.

ggraph(lemur2, layout = "kk") +
  geom_edge_fan() +
  geom_node_point(size = 4) +
  theme_graph()

6.2.6 geom_edge_parallel

geom_edge_parallel does something similar:

ggraph(lemur2, layout = "kk") +
  geom_edge_parallel() +
  geom_node_point(size = 4) +
  theme_graph()

However, we don’t yet see the link towards itself. For this you can use:

6.2.7 geom_edge_loop

ggraph(lemur2, layout = "kk") +
  geom_edge_fan() +
  geom_edge_loop() +
  geom_node_point(size = 4) +
  theme_graph()

6.3 DIY

Try to make something fancy for the Zebras! Data from this wonderful website. 

zebra_graphml <- read.graph("https://raw.githubusercontent.com/bansallab/asnr/master/Networks/Mammalia/zebra_groupmembership_weighted/zebra_sundaresan_interaction_attribute.graphml", format = "graphml")

zebra <- as_tbl_graph(zebra_graphml)