我有节点、边和权重。如何从 R shiny 中的节点、边、权重信息绘制网络

您可以使用将绘图转换为交互式绘图。plotly

plotly::plotly_build:

描述：

此泛型函数创建发送到 plotly.js 进行渲染的列表对象。使用此函数可用于覆盖 ggplotly/plot_ly 提供的默认值或调试渲染错误。

PlotlyR 开源图形库

Plotly 的 R 图形库可制作交互式、出版质量的图形。

plot_object<- plot(g)

plotly_object<-plotly::ggplotly(plot_object)

p<-plotly::plotly_build(plotly_objet)

p$elementId <- NULL

# display the interactive plot: 
p

根据版本的不同，尝试不这样做：p$elementId <- NULLplotly

• github.com/ropensci/plotly # 985：忽略显式提供的 widget ID;Shiny 不使用它们

编辑：

igraph似乎不适用于 Ploty :

• https://github.com/igraph/rigraph/issues/496#issuecomment-990209544

这是另一个受 R-bloggers 启发的答案：R 中的网络图，因为在我阅读的评论之后，不支持 .igraphplotly

请注意，我没有验证网络的逻辑。

您需要运行第一部分和第二部分才能运行有关 networkD3 的第三部分。

有些权重是负面的：我不确定这对网络库来说是件好事，也不确定它们是否为此实现了良好的效果。

flat_dataset <-"r  a b weight
[1,] 1 49 0.13920793
[2,] 2 19 -0.16105155
[3,] 2 29 -0.11896011
[4,] 2 49 0.10574541
[5,] 3 29 -0.13120939
[6,] 3 49 0.17390062
[7,] 4 10 -0.10381197
 snip snip snip
[113,] 49 60 0.11064330
[114,] 49 61 -0.13522543
"

data <-read.table(textConnection(flat_dataset),header=TRUE)

您必须通过完成更改数据：上面的行只是为了有数据来测试它。

print(as_tibble(data))
# A tibble: 114 × 4
   r         a     b weight
   <chr> <int> <int>  <dbl>
 1 [1,]      1    49  0.139
 2 [2,]      2    19 -0.161
 3 [3,]      2    29 -0.119
 4 [4,]      2    49  0.106
 5 [5,]      3    29 -0.131
 6 [6,]      3    49  0.174
 7 [7,]      4    10 -0.104
 8 [8,]      5    33 -0.135
 9 [9,]      6    49  0.175
10 [10,]     7    19  0.187
# … with 104 more rows
# ℹ Use `print(n = ...)` to see more rows

以下已加载的包裹。它们必须先安装。

library(tidyverse)
library(igraph))
library(ggraph)
library(networkD3)
library(tidygraph)

第 1 部分 igraph

文章的一部分，但包含您的数据。数据似乎不完整：几乎所有节点都是孤立的。 如果没有分位数滤波器，它们看起来并不完整。igraph

graph <- igraph::graph_from_data_frame(data, directed = FALSE)
plot(graph)

# Warning message in v(graph):
# “Non-positive edge weight found, ignoring all weights during graph layout.”

data2 <- data[data$weight >= stats::quantile(data$weight, p = 0.98),,]
data2

如果您想要所有数据，则必须在下面代替上述操作：

data2 <- data

第1个图形，但带有分位数滤波器：

data2 %>%
    igraph::graph_from_data_frame(directed = FALSE) %>%
    plot(layout = layout_with_graphopt)

没有分位数滤波器的第一个图形：带有：data2 <- data

第 2 部分和tidygraphggraph

nodes <- c(data$a, data$b) %>% 
    unique() %>%
    # tibble from tidyverse library
    tibble::tibble(label = .) %>%
    tibble::rowid_to_column("id")

# nodes

与标签和 ID 中的重量相同：

nodes <- c(data$a, data$b, data$weight) %>% 
    unique()  %>%
    tibble::tibble(label = .) %>%
    tibble::rowid_to_column("id")

我想（也许很糟糕）对于，那就是：tidygraph::tbl_graph()

• a是from
• b是to
• weight是N

edges <- data2 %>%
    # join the example, but your network seems already made
    # dplyr::left_join(nodes, by = c("Person1"="label")) %>%
    # dplyr from tidyverse package
    dplyr::rename(N = "weight") %>%
    dplyr::rename(from = "a") %>%
    # join the example, but your network seems already made
    # dplyr::left_join(nodes, by = c("Person2"="label")) %>%
    dplyr::rename("to" = "b") %>%
    dplyr::select(from, to, N)



graph_tidy <- tidygraph::tbl_graph(nodes = nodes, edges = edges, directed = FALSE)

graph_tidy

这给了我：

# A tbl_graph: 63 nodes and 3 edges
#
# An unrooted forest with 60 trees
#
# A tibble: 63 × 2
     id label
  <int> <int>
1     1     1
2     2     2
3     3     3
4     4     4
5     5     5
6     6     6
# … with 57 more rows
#
# A tibble: 3 × 3
   from    to     N
  <int> <int> <dbl>
1     7    19 0.187
2    25    35 0.187
3    30    49 0.189

graph_tidy %>%
    dplyr::mutate(Centrality = tidygraph::centrality_authority()) %>%
    ggraph::ggraph(layout = "graphopt") + 
    ggraph::geom_node_point(aes(size=Centrality, colour = label), show.legend = FALSE) +
    ggraph::geom_edge_link(aes(width = N), alpha = 0.8, show.legend = FALSE) + 
    ggraph::scale_edge_width(range = c(0.2, 2)) +
    ggraph::geom_node_text(aes(label = label), repel = TRUE)
# Warning message:
# “Using the `size` aesthetic in this geom was deprecated in ggplot2 3.4.0.
# Please use `linewidth` in the `default_aes` field and elsewhere instead.”
# my comment: linewidth doesn't work there

没有分位数滤波器的相同图形：with（在开头）：data2 <- data

将权重作为 id 和标签的一部分：

第 3 部分networkD3

edges <- edges %>%
        dplyr::mutate(from = from -1, to = to - 1) %>%
        dplyr::mutate(N = N / 200)

# edges

nodes <- nodes %>%
    dplyr::mutate(id=id-1) %>%
    dplyr::mutate(nodesize = igraph::authority.score(graph_tidy)$vector*150)

# nodes

networkD3::forceNetwork(Links = edges, 
             Nodes = nodes, 
             Source = "from", 
             Target = "to", 
             NodeID = "label", 
             Group = "id", 
             opacity = 1,
             fontSize = 14,
             zoom = TRUE, 
             Value = "N", 
             Nodesize = "nodesize", 
             opacityNoHover = TRUE
)
# Warning message:
# “It looks like Source/Target is not zero-indexed. This is required in  JavaScript and so your plot may not render.”
# Nodes is a tbl_df. Converting to a plain data frame.

我看到了一个互动网络.

我可以移动并单击这些点（但在此处打印）。

没有分位数滤波器的相同图形：with（在开头）：data2 <- data

将权重作为 id 和标签的一部分（图表摘录）：

它是一个可缩放的矢量图形 （SVG）：

<div class="forceNetwork html-widget html-fill-item-overflow-hidden html-fill-item html-widget-static-bound" id="htmlwidget-380bfc22386ecfa1b181" style="width:960px;height:500px;">
    <svg width="960" height="500" pointer-events="all">
        <g class="zoom-layer">
            <g>
                <line class="link" style="stroke: rgb(102, 102, 102); opacity: 0.25; stroke-width: 0.00216273px;" x1="410.4125002318474" y1="235.5820858923744" x2="374.73641851106566" y2="252.4766163768353"></line>
...
        </g>
    </svg>
</div>

链接：

• https://www.rdocumentation.org/packages/networkD3/versions/0.4/topics/forceNetwork
• http://www.htmlwidgets.org/showcase_networkD3.html
• http://christophergandrud.github.io/networkD3/