前言

雷达图，Radar（也称为蜘蛛图或网络图）是一种二维图表，用于显示多变量数据。每个变量由从中心延伸出来的辐条（或轴）表示，数据点沿着这些辐条绘制。然后将数据点连接形成一个多边形，通过这种方式可以直观地展示多个变量之间的相对强弱。

• DOI: 10.1038/ncomms15107

在R语言里，作雷达图的主要用fmsb和ggradar这两个R包，前者一般绘制的是多边形雷达图，后者默认绘制的是圆形雷达图。大同小异，大家各自体会～

下面是两个R包的部分演示以及雷达图计算统计量的项目代码分享

下方公众号回复Pradar即可领取项目代码及文件

fmsb

数据格式

有2点需要注意：

• 作图数据必须要有一行最大值和一行最小值，用于雷达图边界划分。
• 列数必须大于2。
  

参数调整

• vlabels：变量标签
• vlcex：控制变量标签的字体大小
• pcol：线条颜色
• pfcol： 填色
• plwd： 行宽
• plty：线型。可以是数值向量 1:6 或字符向量 c(“solid”、“dashed”、“dotted”、“dotdash”、“longdash”、“twodash”)。要删除线，请使用plty = 0或plty = “blank”。
• cglcol：线条颜色
• cglty：线型
• cglwd： 行宽
• axislabcol：轴标签和数字的颜色。默认为“蓝色”。
• caxislabels：用作中心轴上的标签的字符向量。

demo

# Plotting student 1
ggradar(
  df[1, ], 
  values.radar = c("0", "10", "20"),
  grid.min = 0, grid.mid = 10, grid.max = 20
  )

# Define colors and titles
colors <- c("#00AFBB", "#E7B800", "#FC4E07")
titles <- c("Student.1", "Student.2", "Student.3")

# Reduce plot margin using par()
# Split the screen in 3 parts
op <- par(mar = c(1, 1, 1, 1))
par(mfrow = c(1,3))

# Create the radar chart
for(i in 1:3){
  create_beautiful_radarchart(
    data = df[c(1, 2, i+2), ], caxislabels = c(0, 5, 10, 15, 20),
    color = colors[i], title = titles[i]
    )
}
par(op)

# Reduce plot margin using par()
op <- par(mar = c(1, 2, 2, 2))
# Create the radar charts
create_beautiful_radarchart(
  data = df, caxislabels = c(0, 5, 10, 15, 20),
  color = c("#00AFBB", "#E7B800", "#FC4E07")
)
# Add an horizontal legend
legend(
  x = "bottom", legend = rownames(df[-c(1,2),]), horiz = TRUE,
  bty = "n", pch = 20 , col = c("#00AFBB", "#E7B800", "#FC4E07"),
  text.col = "black", cex = 1, pt.cex = 1.5
  )
par(op)

opar <- par() 
# Define settings for plotting in a 3x4 grid, with appropriate margins:
par(mar = rep(0.8,4))
par(mfrow = c(3,4))
# Produce a radar-chart for each student
for (i in 4:nrow(df_scaled2)) {
  radarchart(
    df_scaled2[c(1:3, i), ],
    pfcol = c("#99999980",NA),
    pcol= c(NA,2), plty = 1, plwd = 2,
    title = row.names(df_scaled2)[i]
  )
}
# Restore the standard par() settings
par <- par(opar) 

ggradar

gg系列一贯作风。雷达图大同小异，且直接用表达数据即可，无须另外计算界限值。太过于简单就不展示了

• plot.data：每个个人或团体占一行的数据
• values.radar：在最小、平均和最大网格线上显示的值
• grid.min：绘制最小网格线的值
• grid.mid：绘制平均网格线的值
• grid.max：绘制最大网格线的值

放上教程链接大家跟着来即可：https://www.datanovia.com/en/blog/beautiful-radar-chart-in-r-using-fmsb-and-ggplot-packages/#fmsb-radar-chart

此外还有棒棒糖图：

单样本demo

df2 <- t(exam_scores) %>%
  as.data.frame() %>%
  rownames_to_column("Field")
df2
ggdotchart(
  df2, x = "Field", y = "Student.1",
  add = "segments", sorting = "descending",
  ylab = "Exam Score", title = "Student 1"
  )

多组demo

df3 <- df2 %>%
  select(Field, Student.1, Student.2) %>%
  pivot_longer(
    cols = c(Student.1, Student.2),
    names_to = "student",
    values_to = "value"
  )
head(df3)
ggdotchart(
  df3, x = "Field", y = "value", 
  group = "student", color = "student", palette = "jco",
  add = "segment", position = position_dodge(0.3),
  sorting = "descending"
  )

df4 <- df2 %>%
  select(Field, Student.1, Student.2, Student.3) %>%
  pivot_longer(
    cols = c(Student.1, Student.2, Student.3),
    names_to = "student",
    values_to = "value"
  )
head(df4)
ggdotchart(
  df4, x = "Field", y = "value", 
  group = "student", color = "student", palette = "jco",
  add = "segment", position = position_dodge(0.3),
  sorting = "descending", facet.by = "student",
  rotate = TRUE, legend = "none"
  )

library(GGally)
ggparcoord(
  iris,
  columns = 1:4, groupColumn = 5, order = "anyClass",
  showPoints = TRUE, 
  title = "Parallel Coordinate Plot for the Iris Data",
  alphaLines = 0.3
  ) + 
  theme_bw() +
  theme(legend.position = "top")

实战-给radar添加统计量

这里随便找了个cibersort免疫浸润结果做的，此外还可以换成GSVA基因集评分结果、多基因组间表达等，也一样可以用雷达图。

用fmsb包，输入数据格式是一样的，第一二行为最大值和最小值：

由于计算p值及整合代码篇幅过长，已打包，文末领取～

pdf("radarplot.pdf", width = 9, height = 9)
# par(mar = c(1, 1, 1, 1))
par(family = "sans", cex = 1)
# 绘制雷达图并设置字体颜色
radarchart(data_radar, axistype = 1,
           pcol = colors_border, pfcol = colors_in, plwd = 2,
           cglcol = "grey", cglty = 1, axislabcol = "grey", 
           caxislabels = format(round(seq(0, max(as.numeric(as.character(unlist(data_radar)))), length.out = 5), 4), nsmall = 2), 
           cglwd = 0.8, vlcex = 0.8)

# 添加分组图例
legend(x = 0.9, y = 1.2, legend = rownames(data_radar)[c(3:4)], bty = "n", pch = 16, col = colors_border, text.col = "black", cex = 0.8, pt.cex = 1.5)
dev.off()

下方公众号回复Pradar即可领取项目代码及文件