ggplot2图片自定义与拼图

文摘科技 2024-10-18 09:05 江苏

写在前面

此前我们进行过ggplot2的基础教学(可视化初探(ggplot2))，本次内容我们将进行坐标轴、颜色、主题、字体、图例、标注等方面的修改。更多R语言技巧可参考：

R语言基础学习手册

本文输出图片集锦：

18.1 修改标尺

本次教程除了ggplot2和dplyr包，还需要安装用于数据的ISLR和gapminder；用于改进图形的ggrepel, showtext, patchwork和plotly。

18.1.1 自定义坐标轴

ggplot2中，函数scale_x_* 和 scale_y_* 控制图形的x y轴，其中*指定标尺类型。主要有以下几种：

18.1.1.1 自定义连续型变量的坐标轴

在函数scale_*_continuous 中的一些常用选项包括：

18.1.1.1.1 举例

用数据集mtcars，绘制重量(wt)与汽车的燃油效率(mpg)散点图：

1.简单版

library(ggplot2)
ggplot(data=mtcars,aes(x=wt,y=mpg))+
  geom_point()+
  labs(title = "Fuel efficiency by car weight")

2.修改版

进行如下修改：

1.对于重量(wt)：添加轴标签”Weight(1000lbs)“，标尺范围设为1.5到5.5，使用10个主刻度线，不显示次刻度线；

2.对于每加仑汽油行驶英里数(mpg):添加轴标签”Miles per gallon”，标尺范围设为10-35，将主刻度线设定在10、15、20、25、30和35，以1加仑为单位绘制次刻度线。

library(ggplot2)
ggplot(data=mtcars,aes(x=wt,y=mpg))+
  geom_point()+
  labs(title = "Fuel efficiency by car weight")+
  scale_x_continuous(name = "Weight(1000lbs)",
                     n.breaks=10,
                     limits = c(1.5,5.5),
                     minor_breaks = NULL)+
  scale_y_continuous(name = "Miles per gallon",
                      limits= c(10,35),
                     minor_breaks=seq(10,35,1))

18.1.1.2 自定义分类变量的坐标轴

ggplot2中常用的scale_*_discrete选项如下：

参数	描述
name	标尺名称，与函数labs(x=,y=)等同
breaks	刻度的数字向量
limits	定义标尺及其顺序的值的字符向量
labels	提供标签的字符向量（必须与breaks参数的长度一致）使用labels=abbreviate可以缩短长标签
position	坐标轴位置（y轴的左边/右边，x轴的上/下边）,left or right for y axes, top or bottom for x axes.

18.1.1.2.1 举例

用ISLR包中的Wage数据框进行举例，数据包含2011年收集的美国某地区3000名男性员工的工资和人口统计信息。绘制此数据样本中婚姻状态与受教育程度间的关系图。

1.简单绘图

library(ISLR)

## Warning: 程辑包'ISLR'是用R版本4.3.2 来建造的

## 
## 载入程辑包：'ISLR'

## The following object is masked from 'package:vcd':
## 
##     Hitters

library(ggplot2)
data(Wage,package = "ISLR")
ggplot(data=Wage,aes(maritl,fill=education))+
  geom_bar(position = "fill")+
  labs(title = "Participant Education by Maritlr")

2.修改绘图

针对x轴：x轴名称修改为”Maritl”，将标签前面的编号去掉;

针对y轴：y轴名称修改为”Percent”,y轴使用百分比格式。

library(ISLR)
library(ggplot2)
library(scales)
head(Wage)

##        year age           maritl     race       education             region
## 231655 2006  18 1. Never Married 1. White    1. < HS Grad 2. Middle Atlantic
## 86582  2004  24 1. Never Married 1. White 4. College Grad 2. Middle Atlantic
## 161300 2003  45       2. Married 1. White 3. Some College 2. Middle Atlantic
## 155159 2003  43       2. Married 3. Asian 4. College Grad 2. Middle Atlantic
## 11443  2005  50      4. Divorced 1. White      2. HS Grad 2. Middle Atlantic
## 376662 2008  54       2. Married 1. White 4. College Grad 2. Middle Atlantic
##              jobclass         health health_ins  logwage      wage
## 231655  1. Industrial      1. <=Good      2. No 4.318063  75.04315
## 86582  2. Information 2. >=Very Good      2. No 4.255273  70.47602
## 161300  1. Industrial      1. <=Good     1. Yes 4.875061 130.98218
## 155159 2. Information 2. >=Very Good     1. Yes 5.041393 154.68529
## 11443  2. Information      1. <=Good     1. Yes 4.318063  75.04315
## 376662 2. Information 2. >=Very Good     1. Yes 4.845098 127.11574

data(Wage,package = "ISLR")
ggplot(data=Wage,aes(maritl,fill=education))+
  geom_bar(position = "fill")+
  labs(title = "Participant Education by Maritlr")+
  scale_x_discrete(name="Maritl",
                   labels=c("Never Married","Married","Widowed","Divorced","Separated"))+
  scale_y_continuous(name="Percent",
                   label= percent_format(accuracy=2),
                   n.breaks=10)

注意：垂直轴(y)代表的是数值型变量，因此需要用scale_y_continuous 而不是scale_y_discrete。

18.1.2 自定义颜色

ggplot2中scale_color_* ()函数用于点、线、边界和文本颜色设定， scale_fill_* ()函数用于带面积的形状对象的颜色填充。

常用的设定颜色标尺的函数有：

18.1.2.1 连续型调色板

依然用mtcars数据集举例，绘制了汽车重量(wt)和燃油效率(mpg)的关系图，通过映射发动机排量(disp)到点的颜色来添加第3个变量。

library(ggplot2)
p <- ggplot(mtcars,aes(x=wt,y=mpg,color=disp))+
  geom_point(shape=19,size=3)+
  scale_x_continuous(name = "Weight(1000 1bs.)",
                     n.breaks = 10,
                     minor_breaks = NULL,
                     limits = c(1.5,5.5))+
  scale_y_continuous(name = "Mile per gallon",
                     breaks = seq(10,35,5),
                     minor_breaks = seq(10,35,1),
                     limits = c(10,35))
#设置颜色
p+ggtitle("A.Default color gradient")

p+scale_color_gradient(low = "grey",high = "black")+
  ggtitle("B.Greyscale gradient")

p+scale_color_gradient(low = "red",high = "blue")+
  ggtitle("C.Red-blue color Gradient")

p+scale_color_steps(low = "red",high = "blue")+
  ggtitle("D.Red-blue binned color Gradient")

p+scale_color_steps2(low = "red",mid="white",high = "blue")+
  ggtitle("E.Red-white-blue binned color Gradient")

p+scale_color_viridis_c(direction = -1)+
  ggtitle("F.Viridis color gradient")

可以看出分桶与不分桶的区别是：不分桶的颜色变化是连续的，而分桶的颜色是具有梯度的。

18.1.2.2 分类型调色板

用ISLR包中的Wage数据框进行举例，数据包含2011年收集的美国某地区3000名男性员工的工资和人口统计信息。education是分类变量，映射到离散颜色。

library(ISLR)
library(ggplot2)
library(scales)
b <- ggplot(data=Wage,aes(maritl,fill=education))+
  geom_bar(position = "fill")+
  labs(title = "Participant Education by Maritlr")+
  scale_x_discrete(name="Maritl",
                   labels=c("Never Married","Married","Widowed","Divorced","Separated"))+
  scale_y_continuous(name="Percent",
                   label= percent_format(accuracy=2),
                   n.breaks=10)
#设置颜色
b+ggtitle("A.Default colors")

b+scale_fill_brewer(palette = "Set2")+
  ggtitle("B.ColorBrewer Set2 palette")

b+scale_fill_viridis_d()+
  ggtitle("C.Viridis color scheme")

b+scale_fill_manual(values = c("gold4","orange2","deepskyblue3","brown2","yellowgreen"))+
  ggtitle("D.Manual color scheme")

注意：这里education属于数值型分类变量；另外，注意使用scale_fill_而非scale_color_

18.2 修改主题

ggplot2中 theme()可以用于自定义图形的非数据部分。具体参数可见?theme()

主图元素：

例如：将x y轴的标题设置为14点蓝色字体

library(ggplot2)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point()+
  theme(axis.title = element_text(size = 14,color = "blue"))

18.2.1 预置主题

在ggplot2中，有8个预设主题可供选择。分别是：

theme_gray(): 默认灰色主题，具有简洁的灰色背景和白色网格线。

theme_bw(): 黑白主题，具有白色背景和黑色网格线。

theme_minimal(): 极简主题，具有简洁的白色背景和无网格线

theme_classic(): 经典主题，具有白色背景、黑色xy边框。

theme_void(): 空白主题，没有背景、网格线和边框，只有数据图形。

theme_light(): 亮色主题，具有浅色的背景和网格线。

theme_dark(): 暗色主题，具有深色的背景和网格线。

theme_linedraw(): 线条绘制主题，具有简单的黑色线条和白色背景。

用实例看一下：

library(ggplot2)
p <- ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point()
p+theme_grey()+ggtitle("theme_grey")

p+theme_bw()+ggtitle("theme_bw")

p+theme_minimal()+ggtitle("theme_minimal")

p+theme_classic()+ggtitle("theme_classic")

p+theme_void()+ggtitle("theme_void")

p+theme_light()+ggtitle("theme_light")

p+theme_dark()+ggtitle("theme_dark")

p+theme_linedraw()+ggtitle("theme_linedraw")

18.2.2 自定义字体

18.2.2.1 操作方法

a.下载本地/google字体

A.查找本地字体文件：

findfont <- function(x){
  suppressMessages(require(showtext))
  suppressMessages(require(dplyr))
  filter(font_files(),grepl(x,family,ignore.case=TRUE))%>%
    select(path,file,family,face)
}
findfont("comic")

## Warning: 程辑包'showtext'是用R版本4.3.2 来建造的

## Warning: 程辑包'sysfonts'是用R版本4.3.2 来建造的

## Warning: 程辑包'showtextdb'是用R版本4.3.2 来建造的

## Error in select(., path, file, family, face): 参数没有用(path, file, family, face)

将此文件加载到R，并使用自定义名称”comic”：

font_add("comic",regular = "comic.ttf",
         bold = "comicbd.ttf")

B.goole字体下载方式：

font_add_google("name","family")

name:google字体的名称；family:自定义名称，后续代码中将用此名称来引用该字体。

例如：

font_add_google("Schoolbell","bell")

b.将showtext设为图形输出的设备

showtext_auto()

c.在函数ggplot2的theme()中指定字体

可以使用element_text()指定字体的系列、字形、大小、颜色和方向：theme(* =element_text()), *指与文本相关的theme()参数，包括：

参数	描述
axis.title, axis.title.x, axis.title.y	坐标轴标题
axis.text, axis.text.x, axis.text.y	坐标轴上的刻度线标签
legend.text, legend.title	图例项标签和图例标题
plot.title, plot.subtitle, plot.caption	图例题，副标题和图的标注栏
strip.text, strip.text.x, strip.text.y	分图标签

18.2.2.2 举例

library(extrafont)

## Registering fonts with R

## 
## 载入程辑包：'extrafont'

## The following object is masked from 'package:showtextdb':
## 
##     font_install

library(sysfonts)
font_add("comic",regular = "comic.ttf",
         bold = "comicbd.ttf")

font_add_google("Open Sans","Sans")

showtext_auto()

p <- ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point()+
  labs(title = "Fuel Efficiency by Car Weight")

p+theme(plot.title = element_text(family = "Sans",size=14),
        axis.title = element_text(family = "comic"))

18.2.3 自定义图例

与图例有关的theme()参数：

18.2.3.1 举例

还是用mtcars数据集，wt为x轴，mpg为y轴，根据发动机气缸数量给点选择颜色。并进行图例修改：

将图例放置在图的右上角；添加图例的标题Cylinders”；横向列出图例类别；将图例背景设为浅灰色，并去除主要元素（带颜色的符号）周围的背景；给图例添加白色边框。

library(ggplot2)
p <- ggplot(mtcars,aes(x=wt,y=mpg,color=factor(cyl)))+
  geom_point(size=3)+
  scale_color_discrete(name="Cylinders")+
  labs(title = "Fuel Efficiency for 32 Automobiles",
       x="weight(1000 1bs",
       y="miles per gallon")
p

p+theme(legend.position = c(0.95,0.95),
        legend.justification = c(1,1),
        legend.title = element_text("Cylinders"),
        legend.background = element_rect(fill ="grey",
                                         color = "white"),
        legend.key =element_blank())

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call.graphics(C_text, as.graphicsAnnot(x$label), x$x, x$y, :
## font family 'Cylinders' not found, will use 'sans' instead

## Warning in grid.Call.graphics(C_text, as.graphicsAnnot(x$label), x$x, x$y, :
## font family 'Cylinders' not found, will use 'sans' instead

18.2.4 自定义绘图区

theme()参数可以自定义绘图区，代码清单如下：

18.2.4.1 举例

在这个例子中：主网格线设置为灰色实线，次网格线设置为灰色虚线，带状标签背景设置为白色，并把位置放在上方。

library(ggplot2)
mtcars$am <- factor(mtcars$am,labels = c("Automatic","Manual"))
ggplot(mtcars,aes(x=disp,y=mpg))+
  geom_point(aes(color=factor(cyl)),size=2)+
  geom_smooth(method = "lm",
              formula=y~x+I(x^2),linetype="dotted",se=FALSE)+
  facet_wrap(~am,ncol=2)+
  theme_bw()+
  theme(strip.background = element_rect(fill="white"),
        panel.grid.major = element_line(color = "lightgrey"),
        panel.grid.minor = element_line(color = "lightgrey",
                                        linetype = "dashed"),
        legend.position = "top")

18.3 添加标注

添加标注参数如下：

18.3.1 给数据点添加标签

当观测值以点在图上显示时难以分辨不同的点代表的是哪个观测值，例如mtcars数据中以汽车重量(wt)和行驶英里数(mpg)之间关系图。

library(ggplot2)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point(color="steelblue")

分别使用函数geom_text()和geom_label()对观测点进行添加标签：

library(ggplot2)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point(color="steelblue")+
  geom_text(label=row.names(mtcars))

library(ggplot2)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point(color="steelblue")+
  geom_label(label=row.names(mtcars))

文本重叠严重，下面使用geom_text_repel()和geom_label_repel()函数添加标注：

library(ggplot2)
library(ggrepel)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point(color="steelblue")+
  geom_text_repel(label=row.names(mtcars))

library(ggplot2)
ggplot(mtcars,aes(x=wt,y=mpg))+
  geom_point(color="steelblue")+
  geom_label_repel(label=row.names(mtcars))

结果显示文本重叠的问题被解决了。

18.3.2 给条形添加标签

1.常规条形图

用ISLR包中的Wage数据集举例，画出该数据集中不同婚姻状态的百分比。

先计算出百分比：

library(ISLR)
library(dplyr)
plotdata <- Wage %>%
  group_by(maritl) %>%
  summarise(n=n()) %>%
  mutate(pct=n/sum(n),
         lbls=scales::percent(pct))

plotdata

## # A tibble: 5 × 4
##   maritl               n     pct lbls 
##   <fct>            <int>   <dbl> <chr>
## 1 1. Never Married   648 0.216   21.6%
## 2 2. Married        2074 0.691   69.1%
## 3 3. Widowed          19 0.00633 0.6% 
## 4 4. Divorced        204 0.068   6.8% 
## 5 5. Separated        55 0.0183  1.8%

作图：

注意，这里stat=“identity”是必须，以避免x y 美学投射报错;因为stat默认是count。

柱状图中的geom_text则是主要label(标签内容)、vjust(标签位置)和size(标签大小)的设置。

library(ggplot2)
ggplot(plotdata,aes(x=maritl,y=pct))+
  geom_bar(stat = "identity",fill="steelblue")+
  geom_text(aes(label=lbls),
            vjust=-0.5,size=3)+
  theme_bw()

18.4 图形组合

使用patchwork包进行，首先需要将待组合的单图分别保存为单独的对象并命名，然后直接用｜以及/来拼图，A｜B:将图A和B左右组合，A/B:将图A和B上下组合；并可以使用单括号()创建图形的子组，例如：(A|B)/(C|D)：就是将ABCD图组合为左上A，右上B，左下C和右下D。

举例：

library(ggplot2)
library(patchwork)

## 
## 载入程辑包：'patchwork'

## The following object is masked from 'package:MASS':
## 
##     area

P1 <- ggplot(mtcars,aes(disp,mpg))+
  geom_point()

P2 <- ggplot(mtcars,aes(factor(cyl),mpg))+
  geom_boxplot()

P3 <- ggplot(mtcars,aes(mpg))+
  geom_histogram(bins = 8,color="white")

(P1|P2)/P3