ggraph 是建立在ggplot2一个扩展R包，它的目标是为网络、树和系统树等等表示关系与层次的数据结构进行可视化。ggraph继承了ggplot2的图形语法，建立了面向节点和边的新的图形对象（geom），并添加了布局的概念。可以在ggplot2框架下进行更加灵活美观的网络可视化。


核心概念

ggraph 建立在三个易于理解的核心概念之上：

  1. **布局（Layout）**定义了节点如何放置在绘图上，即将关系结构转换为图形中每个节点的 x 和 y 值。
  2. **节点（Node）**是关系结构中的连接实体。可以使用geom_node_*绘制。
  3. **边（Edge）**是关系结构中实体之间的连接。可以使用geom_edge_*()进行可视化。

ggraph具有自身鲜明的特点：使用ggraph()初始化图形，并指定了布局layout；有用于节点（node）和边（edge）的特有的几何对象（geom）和标度调整方法（scale）。Layout、Node和Edge是ggraph可视化的三个基本要素。


布局（Layout）

布局是绘制特定图形结构时节点的垂直和水平位置。由于布局是节点空间位置的全局规范，因此它跨越了绘图中的所有层，因此应在调用到geom或stat之外进行定义。ggraph()是作为图层的初始化使用。

示例：

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

# 未指定布局，默认布局
ggraph(graph) + 
    geom_edge_link(aes(colour = factor(year))) + 
    geom_node_point()
Using `stress` as default layout

1


ggraph使用

ggraph(graph, layout = "auto", ...)

说明：

  • graph：图形的对象。数据类型：igraph-graph_from_data_frame()，tbl_graph-tbl_graph()，layout_tbl_graph-create_layout(),
  • layout：要创建的布局类型。
  • …：参数传递到布局

ggraph 中可用的不同布局类型均有不同的呈现方式。任何类型的图形数据都可以表示为tbl_graph对象

包含以下layout：

  • auto：默认布局，layout_tbl_graph_auto()
  • igraph：igraph布局算法，layout_tbl_graph_igraph()
    • tree
    • sugiyama
    • bipartite
    • star
    • circle
    • nicely
    • dh
    • gem
    • graphopt
    • grid
    • mds
    • sphere
    • randomly
    • fr
    • kk
    • drl
    • lgl
  • dendrogram：将树状图形中的节点布置为树叶图，layout_tbl_graph_dendrogram()
  • manual：让用户手动指定每个节点的位置，layout_tbl_graph_manual()
  • linear：以线性或圆形排列节点以制作弧形图，layout_tbl_graph_linear()
  • matrix：在对角线上排列节点，从而准备将其用于制作矩阵图，layout_tbl_graph_matrix()
  • treemap：从图形中创建树图，即显示加权层次结构的矩形空间，layout_tbl_graph_treemap()
  • circlepack：将层次结构显示为圆圈内的圆圈，layout_tbl_graph_circlepack()
  • partition：创建冰柱或太阳爆发图，layout_tbl_graph_partition()
  • hive：轴上的定位节点根据节点属性从中心展开，layout_tbl_graph_hive()

对于igraph 定义了一系列不同的布局功能，这些功能都可以通过使用参数指定特定布局的布局提供。为了最大限度地减少键入所有 igraph 算法也可以直接传递到参数中。

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'circle') + 
    geom_edge_link(aes(colour = factor(year))) + 
    geom_node_point()

image-20210317100026413


igraph对象
graph_from_data_frame(d, directed = TRUE, vertices = NULL)

说明：

  • d：包含两列Node关联信息的数据框。
  • directed：是否创建定向图形
  • vertices：包含Node信息的数据框

示例：

library(igraph)

actors <- data.frame(name = c("Alice", "Bob", "Cecil", "David", "Esmeralda"),
                     age = c(48,33,45,34,21),
                     gender = c("F","M","F","M","F"))

relations <- data.frame(from = c("Bob", "Cecil", "Cecil", "David", "David", "Esmeralda"),
                        to = c("Alice", "Bob", "Alice", "Alice", "Bob", "Alice"),
                        same.dept = c(FALSE,FALSE,TRUE,FALSE,FALSE,TRUE),
                        friendship = c(4,5,5,2,1,1), advice=c(4,5,5,4,2,3))

graph <- graph_from_data_frame(d = relations, 
                           directed=TRUE, 
                           vertices=actors)

print(graph, e=TRUE, v=TRUE)

IGRAPH 404a27f DN-- 5 6 -- 
+ attr: name (v/c), age (v/n), gender (v/c), same.dept (e/l),
| friendship (e/n), advice (e/n)
+ edges from 404a27f (vertex names):
[1] Bob      ->Alice Cecil    ->Bob   Cecil    ->Alice David    ->Alice
[5] David    ->Bob   Esmeralda->Alice

tbl_graph对象
tbl_graph(nodes = NULL, edges = NULL, directed = TRUE, node_key = "name")

## S3 method for class 'igraph'
as_tbl_graph(x, ...)
## S3 method for class 'data.frame'
as_tbl_graph(x, directed = TRUE, ...)

说明：

  • nodes：包含Node信息的数据框。
  • edges：包含两列Node关联信息的数据框。
  • directed：是否创建定向图形

示例：

library(tidygraph)

actors <- data.frame(name = c("Alice", "Bob", "Cecil", "David", "Esmeralda"),
                     age = c(48,33,45,34,21),
                     gender = c("F","M","F","M","F"))

relations <- data.frame(from = c("Bob", "Cecil", "Cecil", "David", "David", "Esmeralda"),
                        to = c("Alice", "Bob", "Alice", "Alice", "Bob", "Alice"),
                        same.dept = c(FALSE,FALSE,TRUE,FALSE,FALSE,TRUE),
                        friendship = c(4,5,5,2,1,1), advice=c(4,5,5,4,2,3))

graph <- tbl_graph(nodes = actors, edges = relations)

graph
# A tbl_graph: 5 nodes and 6 edges
#
# A directed acyclic simple graph with 1 component
#
# Node Data: 5 x 3 (active)
  name        age gender
  <fct>     <dbl> <fct> 
1 Alice        48 F     
2 Bob          33 M     
3 Cecil        45 F     
4 David        34 M     
5 Esmeralda    21 F     
#
# Edge Data: 6 x 5
   from    to same.dept friendship advice
  <int> <int> <lgl>          <dbl>  <dbl>
1     1     2 FALSE              4      4
2     3     1 FALSE              5      5
3     3     2 TRUE               5      5
# … with 3 more rows

layout_tbl_graph对象
create_layout(graph, layout, ...)

说明：参考ggraph()

示例：

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

layout <- create_layout(graph, layout = 'circle')

head(layout)
xyname.ggraph.orig_indexcircular.ggraph.index
11.00000000.0000000011FALSE1
20.99597430.0896393122FALSE2
30.98392960.1785568933FALSE3
40.96396290.2660368544FALSE4
50.93623490.3513748255FALSE5
60.90096890.4338837466FALSE6

节点（Node）

网络中的节点是连接的实体。虽然图形中的节点是实体的抽象概念，布局是实体的物理位置，但节点呈现是实体的视觉表现。布局提供了 x 和 y 坐标，这些坐标可用于在绘图窗口中以不同的方式绘制节点。

library(ggraph)
library(igraph)

gr <- graph_from_data_frame(highschool)

ggraph(gr, layout = 'circle') + 
    geom_point(aes(x=x, y=y))

image-20210317103816097

正如上面介绍的，布局返回节点位置和元数据，这用作ggplot2中默认绘图数据。

推荐使用以下方式：

library(ggraph)
library(igraph)

gr <- graph_from_data_frame(highschool)

ggraph(gr, layout = 'circle') + 
    geom_node_point()

geom_node_*

  • geom_node_point

  • geom_node_arc_bar

  • geom_node_circle

  • geom_node_range

  • geom_node_text

  • geom_node_label

  • geom_node_tile

  • geom_node_voronoi

示例1

library(ggraph)
library(igraph)

gr <- graph_from_data_frame(flare$edges, vertices = flare$vertices)

ggraph(gr, layout = 'partition') + 
     geom_node_tile(aes(fill = depth))

image-20210317104927568

示例2

library(ggraph)
library(igraph)

gr <- graph_from_data_frame(flare$edges, vertices = flare$vertices)

ggraph(gr, layout = 'partition', circular = TRUE) + 
     geom_node_arc_bar(aes(fill = depth)) + 
     coord_fixed()

image-20210317105222812

示例3

library(ggraph)
library(igraph)

gr <- graph_from_data_frame(highschool)

ggraph(gr, layout = 'circle') + 
    geom_node_point() +
    geom_node_text(aes(label = name), repel = T)

image-20210317105853965


边（Edge）

边是一个抽象的概念，表示两个节点实体之间的关系。

geom_edge_*

  • geom_edge_arc
  • geom_edge_bend
  • geom_edge_density
  • geom_edge_diagonal
  • geom_edge_elbow
  • geom_edge_fan
  • geom_edge_hive
  • geom_edge_loop
  • geom_edge_parallel
  • geom_edge_point
  • geom_edge_span
  • geom_edge_tile

示例1

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_link(aes(colour = factor(year)))

image-20210317124032312

示例2

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_fan(aes(colour = factor(year)))

image-20210317123949515

示例3

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_parallel(aes(colour = factor(year)))

image-20210317124136863

示例4

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_density(aes(fill = factor(year)))

image-20210317124333489

示例5

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_arc(aes(colour = factor(year)))

image-20210317124436403


箭头

示例

library(ggraph)
library(igraph)

graph <- graph_from_data_frame(highschool)

ggraph(graph, layout = 'kk') + 
    geom_edge_link(aes(colour = factor(year)), 
                  arrow = arrow(length = unit(4, 'mm')),
                  start_cap = circle(3, 'mm'),
                  end_cap = circle(3, 'mm')) +
    geom_node_point(size = 5)

image-20210317125315993


其他函数

  1. facet_nodes()、facet_edges()、facet_graph()：分片。
  2. theme_graph()：ggraph的主题。
  3. scale_edge_*：边的标度

最终示例

library(ggraph)
library(igraph)

actors <- data.frame(name = c("Alice", "Bob", "Cecil", "David", "Esmeralda"),
                     age = c(48,33,45,34,21),
                     gender = c("F","M","F","M","F"))

relations <- data.frame(from = c("Bob", "Cecil", "Cecil", "David", "David", "Esmeralda"),
                        to = c("Alice", "Bob", "Alice", "Alice", "Bob", "Alice"),
                        same.dept = c(FALSE,FALSE,TRUE,FALSE,FALSE,TRUE),
                        friendship = c(4,5,5,2,1,1), advice=c(4,5,5,4,2,3))

graph <- graph_from_data_frame(d = relations, 
                           directed=TRUE, 
                           vertices=actors)
                    
ggraph(graph, layout = 'kk') + 
    geom_edge_link(mapping = aes(edge_width = friendship, edge_color = advice),
                   arrow = arrow(length = unit(4, 'mm')),
                   start_cap = circle(3, 'mm'),
                   end_cap = circle(3, 'mm')) +
    geom_node_point(size = 5,
                    mapping = aes(colour = gender),
                    alpha = 0.5) +
    geom_node_text(mapping = aes(label=name),repel = T,size = 4) +
    scale_edge_width_continuous(range = c(0.5,0.9)) +
    scale_edge_color_continuous(low = "orange",high = "red")+
    scale_color_manual(values = c("red", "blue"))+
    theme_graph()

image-20210317150436082