The Grammar of Graphics

library(ggplot2)
ggplot(data = mpg,
    mapping = aes(x = displ,
                  y = hwy,
                  colour = class)) +
    geom_point()+
    coord_equal(ratio = 1/5)+
    facet_wrap(.~cyl,nrow = 1)+
    scale_color_viridis_d()+
    theme_bw()+
    theme(panel.grid.major = element_blank(),
          plot.title = element_text(size = 16),
          axis.line = element_line(size = 1),
          panel.background = element_rect(fill = 'transparent', colour = 'black'),
          plot.margin = margin(5,5, 5,5))+
    labs(title = 'A',
         colour = 'Class of car')

基于 ggplot2 可视化的入门教程

Create Elegant Data Visualisations Using the Grammar of Graphics

2024-01-18

Kangguo Li

School of Public Health, Xiamen University

Press Space for next page

Introduction

ggplot2 is a system for declaratively creating graphics, based on The Grammar of Graphics. You provide the data, tell ggplot2 how to map variables to aesthetics, what graphical primitives to use, and it takes care of the details.

ggplot2 是一个基于图形语法声明式创建图形的系统。您提供数据，告诉 ggplot2 如何将变量映射到美学，使用什么图形代码，它会处理细节。

The tidyverse is an opinionated collection of R packages designed for data science. All packages share an underlying design philosophy, grammar, and data structures.

ggplot2 是 tidyverse 的一部分，也就是说我们安装 tidyverse 的时候，ggplot2 也会被安装。

Installation

install.packages("tidyverse") # Install tidyverse
library(tidyverse) # Load tidyverse

Source: The Tidyverse · Teach Data Science. (2024, January 18). Retrieved from https://teachdatascience.com/tidyverse

What is a Plot?

🕵️‍♀ Question: 大家都知道哪些统计图形？

运行这个代码，你会发现，这个图形是由一系列的点组成的，这些点的位置是由 x 和 y 决定的，而点的颜色是由 class 决定的。

ggplot(data = mpg,

       mapping = aes(x = displ,
                     y = hwy,
                     colour = class)) +

  geom_point()

🕵️‍♀ Question: 这一个统计图由哪些元素构成的？

- Data(数据集)

- Aesthetics(美学映射)

- Geometry(几何对象)

Source: Create Elegant Data Visualisations Using the Grammar of Graphics. (2023, December 18). Retrieved from https://ggplot2.tidyverse.org

The Grammar of Graphics

library(ggplot2)
ggplot(data = mpg,
    mapping = aes(x = displ,
                  y = hwy,
                  colour = class)) +
    geom_point()+
    coord_equal(ratio = 1/5)+
    facet_wrap(.~cyl,nrow = 1)+
    scale_color_viridis_d()+
    theme_bw()+
    theme(panel.grid.major = element_blank(),
          plot.title = element_text(size = 16),
          axis.line = element_line(size = 1),
          panel.background = element_rect(fill = 'transparent', colour = 'black'),
          plot.margin = margin(5,5, 5,5))+
    labs(title = 'A',
         colour = 'Class of car')

theme(主题): theme_bw()
Space(空间): coord_cartesian()
Statistical models(统计变换): stat_smooth()
Facets(分面): facet_wrap()
Geometry(几何对象): geom_point()
Aesthetics(美学映射): x, y, color
Data(数据集): mpg

Source: Data Visualization Cheatsheet · RStudio. (2024, January 18). Retrieved from https://rstudio.com/resources/cheatsheets/

Data

数据是统计图形的核心，作图之前一定要先读懂数据。

长数据

个体	变量	值
1	x	1
1	y	2
2	x	3
2	y	4

宽数据

个体	x	y
1	1	2
2	3	4

🕵️‍♀ Question: 长数据还是宽数据？

一般选择长数据，因为长数据更容易处理。

Data transformation

数据的长宽转换

library(tidyverse)
data <- tibble(
  x = 1:4,
  y = 1:4,
  z = 1:4
)
data

      x     y     z
  <int> <int> <int>
1     1     1     1
2     2     2     2
3     3     3     3
4     4     4     4

data |>
    pivot_longer(cols = everything())

   name  value
   <chr> <int>
 1 x         1
 2 y         1
 3 z         1
 4 x         2
 5 y         2
 6 z         2
 7 x         3

data |>
    pivot_longer(cols = c(y, z),
                 names_to = 'name',
                 values_to = 'value')

       x name  value
   <int> <chr> <int>
 1     1 y         1
 2     1 z         1
 3     2 y         2
 4     2 z         2
 5     3 y         3

看起来很简单对吧？但是这个函数的参数有很多，你可以通过?pivot_longer查看。

Aesthetics

美学映射是指将数据集中的变量映射到图形的视觉属性上，比如颜色、形状、大小等。

x
y
fill
color
shape
linetype
size
alpha
group

这些并不是固定的，你可以自己定义。

如果需要将变量映射到美学上，需要在aes()中定义。

对于支持的美学映射，可以通过?aes或者?geom_xxxx查看。

library(ggplot2)
ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy,
                             colour = class))+
    coord_equal(ratio = 1/5)+
    facet_wrap(.~cyl,nrow = 1)+
    scale_color_viridis_d()+
    theme_bw()+
    theme(panel.grid.major = element_blank(),
          plot.title = element_text(size = 16),
          axis.line = element_line(size = 1),
          panel.background = element_rect(fill = 'transparent', colour = 'black'),
          plot.margin = margin(5,5, 5,5))+
    labs(title = 'A',
         colour = 'Class of car')

🕵️‍♀ Question: 哪些变量被映射到了美学上？

displ, hwy, class

Geometry

几何对象是指图形的形状，比如点、线、面等。

也就是说我们需要通过geom_xxxx()来定义图形的形状，如散点图、折线图、柱状图等。

🕵️‍♀ Question: 回忆下一些统计学知识，你知道哪些图形？它们分别适用于什么变量？

如：散点图：两个连续性/定量变量

我们通常将变量分为几类

定性变量，如性别、学历等。
- 有序定性变量，如学历。
- 无序定性变量，如性别。
定量变量，如年龄、身高等。

One variable

定量变量

定性变量

直方图

ggplot(data = mpg) +
    geom_histogram(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_freqpoly(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_density(mapping = aes(x = displ))

频率图

ggplot(data = mpg) +
    geom_histogram(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_freqpoly(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_density(mapping = aes(x = displ))

核密度图

ggplot(data = mpg) +
    geom_histogram(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_freqpoly(mapping = aes(x = displ))

ggplot(data = mpg) +
    geom_density(mapping = aes(x = displ))

Two variables

定量+定量

常用的只有 geom_point, geom_line, geom_smooth, geom_area

散点图

ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy))

ggplot(data = mpg) +
    geom_line(mapping = aes(x = displ,
                            y = hwy))

ggplot(data = mpg) +
    geom_smooth(mapping = aes(x = displ,
                              y = hwy))

ggplot(data = mpg) +
    geom_area(mapping = aes(x = displ,
                            y = hwy))

折线图

ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy))

ggplot(data = mpg) +
    geom_line(mapping = aes(x = displ,
                            y = hwy))

ggplot(data = mpg) +
    geom_smooth(mapping = aes(x = displ,
                              y = hwy))

ggplot(data = mpg) +
    geom_area(mapping = aes(x = displ,
                            y = hwy))

平滑曲线

ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy))

ggplot(data = mpg) +
    geom_line(mapping = aes(x = displ,
                            y = hwy))

ggplot(data = mpg) +
    geom_smooth(mapping = aes(x = displ,
                              y = hwy))

ggplot(data = mpg) +
    geom_area(mapping = aes(x = displ,
                            y = hwy))

面积图

ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy))

ggplot(data = mpg) +
    geom_line(mapping = aes(x = displ,
                            y = hwy))

ggplot(data = mpg) +
    geom_smooth(mapping = aes(x = displ,
                              y = hwy))

ggplot(data = mpg) +
    geom_area(mapping = aes(x = displ,
                            y = hwy))

定量+定性

柱状图

ggplot(data = mpg) +
    geom_col(mapping = aes(x = class,
                           y = cyl))

ggplot(data = mpg) +
    geom_boxplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_dotplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_violin(mapping = aes(x = class,
                              y = hwy))

箱式图

ggplot(data = mpg) +
    geom_col(mapping = aes(x = class,
                           y = cyl))

ggplot(data = mpg) +
    geom_boxplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_dotplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_violin(mapping = aes(x = class,
                              y = hwy))

点图

ggplot(data = mpg) +
    geom_col(mapping = aes(x = class,
                           y = cyl))

ggplot(data = mpg) +
    geom_boxplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_dotplot(mapping = aes(x = class,
                               y = hwy,
                               fill = class),
                 binaxis = 'y',
                 stackdir = 'center')

ggplot(data = mpg) +
    geom_violin(mapping = aes(x = class,
                              y = hwy))

小提琴

ggplot(data = mpg) +
    geom_col(mapping = aes(x = class,
                           y = cyl))

ggplot(data = mpg) +
    geom_boxplot(mapping = aes(x = class,
                               y = hwy))

ggplot(data = mpg) +
    geom_dotplot(mapping = aes(x = class,
                               y = hwy,
                               fill = class),
                 binaxis = 'y',
                 stackdir = 'center')

ggplot(data = mpg) +
    geom_violin(mapping = aes(x = class,
                              y = hwy))

Two variables

定性+定性

特殊情况地图

常用使用 sf 提供的 geom_sf.

热图

相信你已经掌握了这些简单的图形，接下来我们来看看一些高级的用法。

Three variables

定性+定性+定性
定性+定性+定量
定性+定量+定量
定量+定量+定量

Solution 1: 热图类

Source: China CDC Weekly

热图类 Test data

data <- data.frame(
  x = rep(1:10, 10),
  y = rep(1:10, each = 10),
  z = rnorm(100)
)

ggplot(data = data) +
    geom_contour_filled(mapping = aes(x = x,
                                       y = y,
                                       z = z))

ggplot(data = data) +
    geom_contour(mapping = aes(x = x,
                               y = y,
                               z = z))

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))

ggplot(data = data) +
    geom_raster(mapping = aes(x = x,
                              y = y,
                              fill = z))

热图类

data <- data.frame(
  x = rep(1:10, 10),
  y = rep(1:10, each = 10),
  z = rnorm(100)
)

ggplot(data = data) +
    geom_contour_filled(mapping = aes(x = x,
                                       y = y,
                                       z = z))

ggplot(data = data) +
    geom_contour(mapping = aes(x = x,
                               y = y,
                               z = z))

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))

ggplot(data = data) +
    geom_raster(mapping = aes(x = x,
                              y = y,
                              fill = z))

热图类

data <- data.frame(
  x = rep(1:10, 10),
  y = rep(1:10, each = 10),
  z = rnorm(100)
)

ggplot(data = data) +
    geom_contour_filled(mapping = aes(x = x,
                                       y = y,
                                       z = z))

ggplot(data = data) +
    geom_contour(mapping = aes(x = x,
                               y = y,
                               z = z))

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))

ggplot(data = data) +
    geom_raster(mapping = aes(x = x,
                              y = y,
                              fill = z))

数据集很大使用geom_raster; 数据集较小geom_tile

Three variables

定性+定性+定性
定性+定性+定量
定性+定量+定量
定量+定量+定量

Solution 1: 热图类

Source: China CDC Weekly

Solution 2: 3D plot

Source: Nature

非必要不使用

Solution 3: Facet

Source: Nature

除了热图这种比较常用的以外，3D 图标也会经常用到，但是 3D 图标的缺点是不够直观，而且不够美观。不建议大家使用，也就是说非必要不适用。

Facets

一个图拆分为多个子图，每个子图展示不同的数据。

Why facet?

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_wrap(.~class,nrow = 2)

facet_grid

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_grid(.~cyl)

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_grid(cyl~.)

facet_grid

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_grid(cyl~.)

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_grid(cyl~.,
             scales = 'free')

facet_wrap

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_grid(cyl~.,
             scales = 'free')

ggplot(data = mpg) +
  geom_point(mapping = aes(x = displ,
                           y = hwy))+
  facet_wrap(cyl~.,
             scales = 'free')

summary

如果需要任意堆积，可以使用facet_wrap()。如果要按照某个变量堆积，可以使用facet_grid()。

对于facet_wrap()，可以使用nrow和ncol来控制子图的行数和列数。

更多信息可以通过?facet_wrap和?facet_grid查看。

Three variables

定性+定性+定性
定性+定性+定量
定性+定量+定量
定量+定量+定量

Solution 1: 热图类

Source: China CDC Weekly

Solution 2: 3D plot

Source: Nature

Solution 3: Facet

Source: Nature

Solution 4: Aesthetics: color, shape, size…

Source: Nature

Statistical models

我更倾向于先把数据处理好，再作图。

Space

Coordinate systems

coord_cartesian: 用于裁剪图片修正显示范围，常用于制作字母图。
coord_flip: 用于翻转坐标轴。
coord_fixed: 用于固定坐标轴比例，常用于绘制热图和地图。
coord_polar 和 coord_quickmap：使用较少。

Theme

设置背景和坐标轴

library(ggplot2)
ggplot(data = mpg) +
    geom_point(mapping = aes(x = displ,
                             y = hwy,
                             colour = class))+
    coord_equal(ratio = 1/5)+
    facet_wrap(.~cyl,nrow = 1)+
    scale_color_viridis_d()+
    theme_bw()+
    theme(panel.grid.major = element_blank(),
          plot.title = element_text(size = 16),
          axis.line = element_line(size = 1),
          panel.background = element_rect(fill = 'transparent', colour = 'black'),
          plot.margin = margin(5,5, 5,5))+
    labs(title = 'A',
         colour = 'Class of car')

主题模板 + 主题自定义

🕵️‍♀ Question: 这两部分能否调转过来？

What else?

Scales

Scales of axis

根据变量类型设置坐标轴 (以 x 轴为例)

Variable type	R class	Scale function
定量变量	numeric/integer	scale_x_continuous()
定性变量	factor/character	scale_x_discrete()
日期/时间	date/POSIXct	scale_x_date()

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))+
    scale_x_continuous(expand = c(0,0))+
    scale_y_continuous(expand = c(0,0))

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))+
    scale_x_continuous(expand = c(0,0))+
    scale_y_continuous(expand = c(0,0))

Scales of guides

根据变量类型设置图例 (以 fill 轴为例)

Variable type	R class	Scale function
定量变量	numeric/integer	scale_fill_distiller()
		scale_fill_gradient()
		scale_fill_gradient2()
		scale_fill_gradientn()
定性变量	factor/character	scale_fill_brewer()
		scale_fill_manual()

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))+
    scale_x_continuous(expand = c(0,0))+
    scale_y_continuous(expand = c(0,0))

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))+
    scale_x_continuous(expand = c(0,0))+
    scale_y_continuous(expand = c(0,0))+
    scale_fill_distiller(palette = 'Blues', direction = 1)

Labels and legends

Labels

ggplot(data = data) +
    geom_tile(mapping = aes(x = x,
                            y = y,
                            fill = z))+
    scale_x_continuous(expand = c(0,0))+
    scale_y_continuous(expand = c(0,0))+
    scale_fill_distiller(palette = 'Blues', direction = 1)+
    labs(title = 'A',
         subtitle = NULL,
         x = 'X label',
         y = 'Y label',
         fill = 'Z label')

guides

ggplot(data = data) +
  geom_tile(mapping = aes(x = x, y = y, fill = z)) +
  scale_x_continuous(expand = c(0,0)) +
  scale_y_continuous(expand = c(0,0)) +
  scale_fill_distiller(palette = 'Blues', direction = 1) +
  labs(title = 'A',
       subtitle = NULL,
       x = 'X label',
       y = 'Y label',
       fill = 'Z label') +
  guides(fill = guide_colorbar(
      barwidth = 1,
      barheight = 15
  ))

Zooming

Save plot

# png
ggsave('plot.png', width = 10, height = 10, dpi = 300)
# pdf
ggsave('plot.pdf', width = 10, height = 10, device = cairo_pdf)
# pdf & 中文
ggsave('plot.pdf', width = 10, height = 10, device = cairo_pdf, family = 'GB1')

Using esquisse

Epidemiologist Toolbox

How to learn?

前人的工作+你的工作=最佳的统计图形

天才是极少数的，一个统计图形的设计需要经过大量的尝试，所以我们需要学习前人的工作。

Source: Science Advances

Source: Nature

Learn More

PDF download GitHub Repository My Resume

The Grammar of Graphics

基于 ggplot2 可视化的入门教程

Create Elegant Data Visualisations Using the Grammar of Graphics

2024-01-18

Kangguo Li

School of Public Health, Xiamen University

Table of Contents

Introduction

What is a Plot?

The Grammar of Graphics

Data

Data transformation

Aesthetics

Geometry

One variable

直方图

频率图

核密度图

Two variables

散点图

折线图

平滑曲线

面积图

柱状图

箱式图

点图

小提琴

Two variables

Three variables

热图类 Test data

热图类

热图类

Three variables

Facets

Why facet?

facet_grid

facet_grid

facet_wrap

summary

Three variables

Statistical models

Space

Theme

What else?

Scales

Scales of axis

Scales of guides

Labels and legends

Labels

guides

Zooming

Save plot

Using esquisse

How to learn?

Learn More