我正试图以d3Network的R端口为例,创建一个详细说明的Sankey Plot(如下所述:https://christophergandrud.github.io/networkD3/)。我加载以下示例"能量"数据集:
# Load energy projection data
URL <- paste0("https://cdn.rawgit.com/christophergandrud/networkD3/",
"master/JSONdata/energy.json")
Energy <- jsonlite::fromJSON(URL)
导入"能源"数据集会生成两个新的数据帧;节点和链接。查看链接数据可以发现以下格式:
head(Energy$links)
source target value
1 0 1 124.729
2 1 2 0.597
3 1 3 26.862
4 1 4 280.322
5 1 5 81.144
6 6 2 35.000
"源"列表示源节点,"目标"列表示目标节点,而"值"列表示每个单独链接的值。
尽管这在概念上相当简单,但我在获得与Energy$links
data.frame格式相同的数据集时遇到了巨大的困难。我已经能够获得以下格式的数据,但我对如何进一步转换它一无所知:
head(sampleSankeyData, n = 10L)
clientID node1
<int> <chr>
1 23969 1 Community Services
2 39199 1 Youth Justice
3 23595 1 Mental Health
4 15867 1 Community Services
5 18295 3 Housing
6 18295 2 Housing
7 18295 1 Community Services
8 18295 4 Housing
9 15253 1 Housing
10 27839 1 Community Services
我想做的是为每个链接聚合唯一客户端的数量。例如,在上述数据子集中,由于客户18295,"1社区服务"到"2住房"的链接值应为1("2住房"到"3住房"以及"3住房"到"4住房"的链接值也应为1)。因此,我希望能够获得与Sankey图示例中的Energy$links
相同格式的数据。
试试这个:
library(tidyverse)
library(stringr)
df <- tribble(
~number, ~clientID, ~node1,
1 , 23969, '1 Community Services',
2 , 39199, '1 Youth Justice',
3 , 23595, '1 Mental Health',
4 , 15867, '1 Community Services',
5 , 18295, '3 Housing',
6 , 18295, '2 Housing',
7 , 18295, '1 Community Services',
8 , 18295, '4 Housing',
9 , 15253, '1 Housing',
10, 27839, '1 Community Services')
df2 <- mutate(df, step=as.numeric(str_sub(node1, end=1))) %>%
spread(step, node1, sep='_') %>%
group_by(clientID) %>%
summarise(step1 = sort(unique(step_1))[1],
step2 = sort(unique(step_2))[1],
step3 = sort(unique(step_3))[1],
step4 = sort(unique(step_4))[1])
df3 <- bind_rows(select(df2,1,source=2,target=3),
select(df2,1,source=3,target=4),
select(df2,1,source=4,target=5)) %>%
group_by(source, target) %>%
summarise(clients=n())
并将其用于CCD_ 3。。。
links <- df3 %>%
dplyr::ungroup() %>% # ungroup just to be safe
dplyr::filter(!is.na(source) & !is.na(target)) # remove lines without a link
# build the nodes data frame based on nodes in your links data frame
nodeFactors <- factor(sort(unique(c(links$source, links$target))))
nodes <- data.frame(name = nodeFactors)
# convert the source and target values to the index of the matching node in the
# nodes data frame
links$source <- match(links$source, levels(nodeFactors)) - 1
links$target <- match(links$target, levels(nodeFactors)) - 1
# plot
library(networkD3)
sankeyNetwork(Links = links, Nodes = nodes, Source = 'source',
Target = 'target', Value = 'clients', NodeID = 'name')