欢迎关注R语言数据分析指南
❝之前介绍了如何使用「networkD3」包来绘制交互式桑基图，本节再来介绍如何使用「ggsankey」绘制有多个分类变量的桑基图。下面就来通过一个小案例介绍 ❞
加载R包
代码语言：javascript复制library(tidyverse)
# remotes::install_local("ggsankey-main.zip",upgrade = F,dependencies = T)
library(ggsankey)
library(ggtext)
数据清洗
代码语言：javascript复制frogs <- read_csv("frogs.txt") %>% # 从"frogs.txt"文件中读取数据，并将结果保存到变量frogs中
arrange(Ordinal) %>% # 按照Ordinal列的值对数据进行排序
mutate(SurveyDate = as.Date(SurveyDate, format = "%m/%d/%Y"), # 将SurveyDate列转换为日期格式
Gender = if_else(Female == 0, "Male", "Female")) %>% # 根据Female列的值判断Gender
group_by(Subsite) %>% # 按照Subsite列进行分组
mutate(total = n()) %>% # 计算每个分组中的观测数量，并将结果保存到total列中
ungroup() %>% # 取消分组
mutate(Subsite2 = glue::glue("<b>{Subsite}</b> ({total} frogs)"))
构建数据格式
代码语言：javascript复制dt2 <- make_long(frogs, HabType, Water, Type, Structure, Substrate, value = "Subsite2") # 使用make_long()函数对数据进行重塑，结果保存到dt2变量中
数据可视化
代码语言：javascript复制ggplot(dt2,aes(x = x, next_x = next_x,
node = node, next_node = next_node)) # 创建一个ggplot对象，并指定数据和映射关系
geom_sankey(flow.alpha = 0.5,node.fill = "#f5ccae", # 添加sankey流程图层，设置外观参数
flow.fill = "#aeb4f5",node.color = "#aef5ef",
flow.color = "#aef5ef",flow.size = 0.25,node.size = 0.25)
geom_sankey_text(aes(label = node), size = 3) # 添加sankey流程图文本标签层
facet_wrap(vars(value), scales = "free",ncol = 3) # 按照value列进行分面，每行显示3个图形
coord_cartesian(expand = TRUE, clip = "off") # 设置坐标系，展开绘图范围并关闭裁剪
theme_minimal() # 使用最简主题
theme(legend.position = "bottom", # 设置主题参数
panel.grid = element_blank(),
panel.spacing = unit(1,"lines"),
plot.background = element_rect(fill = "#f5faf8", color = "#f5faf8"),
axis.text.y = element_blank(),
axis.title.x = element_blank(),
strip.text = element_markdown(margin = margin(t = 10, b = 5),face="bold",color="black"))
