10 R语言绘制桑基图详细教程
本期主要介绍常用桑基图（流量图）绘制包ggsankeyfier和相关代码教程
桑基图也叫流量图，主要用来处理多个分类变量的分布情况。可视化展示占比比例等，美观容易
10.3 R包使用
桑基图本质上就是不同分类变量的流量图，可视化的第一步就是数据整洁化，常用数据宽变长。一列是变量，一列是流量值，在ggsankeyfier使用过程中重要的就是用pivot_stages_longer来做数据清洗
10.3.1 数据准备
## first get a data set with a wide format:
data(ecosystem_services)
data <- ecosystem_services
## pivot to long format for plotting:
data_neat <-
data |>
subset(RCSES > 0.005) |>
pivot_stages_longer(
c("activity_realm", "biotic_realm", "service_section"),
"RCSES", "service_section"
)
# pivot宽变长前
head(data, 5)
## # A tibble: 5 × 8
## activity_type activity_realm pressure_cat biotic_group biotic_realm
## <chr> <chr> <chr> <chr> <chr>
## 1 Agriculture & Forestry Land-based Biological di… Zooplankton Plankton
## 2 Agriculture & Forestry Land-based Biological di… Zooplankton Plankton
## 3 Agriculture & Forestry Land-based Chemical chan… Zooplankton Plankton
## 4 Agriculture & Forestry Land-based Chemical chan… Zooplankton Plankton
## 5 Agriculture & Forestry Land-based Physical chan… Zooplankton Plankton
## # ℹ 3 more variables: service_division <chr>, service_section <chr>,
## # RCSES <dbl># 宽变长后
head(data_neat, 5)
## # A tibble: 5 × 6
## service_section RCSES edge_id connector node stage
## <fct> <dbl> <int> <chr> <fct> <fct>
## 1 Provisioning 0.0123 1 from Land-based activity_realm
## 2 Provisioning 0.0123 1 to Plant biotic_realm
## 3 Provisioning 0.0133 2 from Land-based activity_realm
## 4 Provisioning 0.0133 2 to Fish & Cephalopods biotic_realm
## 5 Cultural 0.0127 3 from Land-based activity_realm10.4 总绘图代码
data_neat |>
ggplot(
mapping = aes(
x = stage, y = RCSES,
group = node, edge_id = edge_id, connector = connector
)
) +
geom_sankeyedge(aes(alpha = RCSES, fill = service_section),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.01
)
) +
geom_sankeynode(aes(fill = node, color = "grey"),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.11
),
show.legend = FALSE
) +
geom_text(
data = data_neat,
aes(label = node), stat = "sankeynode",
position = position_sankey(v_space = "auto", order = "ascending", nudge_x = 0),
hjust = 0.5, size = 3.5, fontface = "bold", color = "black"
)+
guides(
alpha = FALSE,
colour = FALSE,
fill = guide_legend(position = "bottom", nrow = 1)
) +
scale_fill_discrete(
breaks = unique(data_neat$service_section)[1:3] # 只在图例中显示前三个键
)+
theme_minimal_hgrid() +
theme( # 去除周围外边距
panel.spacing = unit(0, "cm"),
panel.grid.major.y = element_blank(),
axis.text.y = element_blank(),
axis.title = element_blank(),
legend.key.spacing.x = unit(60, "pt")
) 
10.5 绘图代码拆解
接下来我们拆解上边一大段代码，一步步的跑出来就会感觉会好很多。数据的预先准备至关重要，看数据准备部分的代码就行了
10.5.1 第一步我们先把sankeyedge做出来，就是中间的流量
看到我们设置了一个至关重要的的参数position：这个参数决定了sankeyedge的位置，我们设置为position_sankey(order ="ascending",v_space="auto",width = 0.01)，其中order是设置sankeyedge的方向，ascending是从左到右，descending是从右到左，v_space是设置sankeyedge的间距，auto是自动设置，width是设置sankeyedge的宽度。
data_neat |>
ggplot(
mapping = aes(
x = stage, y = RCSES,
group = node, edge_id = edge_id, connector = connector, fill = stage
)
) +
geom_sankeyedge(aes(alpha = RCSES, fill = service_section),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.01
)
)
10.5.2 很好，下一部我们把node给加上，就是各个小方块
设置node的填充色是node，然后边框设置成灰色
data_neat |>
ggplot(
mapping = aes(
x = stage, y = RCSES,
group = node, edge_id = edge_id, connector = connector
)
) +
geom_sankeyedge(aes(alpha = RCSES, fill = service_section),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.01
)
) +
geom_sankeynode(aes(fill = node, color = "grey"),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.1
)
)
10.5.3 然后我们添加文字到图形里
注意添加 stat = "sankeynode"
data_neat |>
ggplot(
mapping = aes(
x = stage, y = RCSES,
group = node, edge_id = edge_id, connector = connector
)
) +
geom_sankeyedge(aes(alpha = RCSES, fill = service_section),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.01
)
) +
geom_sankeynode(aes(fill = node, color = "grey"),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.1
)
) +
geom_text(
data = data_neat,
aes(label = node), stat = "sankeynode",
position = position_sankey(v_space = "auto", order = "ascending", nudge_x = 0),
hjust = 0.5, size = 3.5, fontface = "bold", color = "black"
)
10.5.4 最后阶段我们修改一下图例，改改主题
我们的主要变量是service_section（geom_sankeyedge的fill映射），我们在这里就设置这个作为图例，其它都删掉。这一部分说起来简单，其实挺麻烦的。
data_neat |>
ggplot(
mapping = aes(
x = stage, y = RCSES,
group = node, edge_id = edge_id, connector = connector
)
) +
geom_sankeyedge(aes(alpha = RCSES, fill = service_section),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.01
)
) +
geom_sankeynode(aes(fill = node, color = "grey"),
position = position_sankey(
order = "ascending", v_space = "auto",
width = 0.11
),
show.legend = FALSE
) +
geom_text(
data = data_neat,
aes(label = node), stat = "sankeynode",
position = position_sankey(v_space = "auto", order = "ascending", nudge_x = 0),
hjust = 0.5, size = 3.5, fontface = "bold", color = "black"
)+
guides(
alpha = FALSE,
colour = FALSE,
fill = guide_legend(position = "bottom", nrow = 1)
) +
scale_fill_discrete(
breaks = unique(data_neat$service_section)[1:3] # 只在图例中显示前三个键
)+
theme_minimal_hgrid() +
theme( # 去除周围外边距
panel.spacing = unit(0, "cm"),
panel.grid.major.y = element_blank(),
axis.text.y = element_blank(),
axis.title = element_blank(),
legend.key.spacing.x = unit(60, "pt")
) 