利用R进行数据可视化——第九章重中之重之自定义图形

图形默认值适用于快速的数据探索，但是当你希望将结果发布到博客、论文、文章或海报时，您可能需要自定义结果。定制可以提高图形的清晰度和吸引力。

本章将描述如何自定义图形的轴、网格线、颜色、字体、标签和图例。它还描述了如何添加注释(文本和行)。

---

9.1 坐标轴

x轴和y轴表示连续、分类或日期值。您可以使用下面的函数修改默认的刻度和标签。

9.1.1 数值轴

使用scale_x_continuous或scale_y_continuous函数修改定量轴。

选项包括

• breaks-位置限制的数值向量
• limits-带有刻度的最小值和最大值的数值向量

library(pacman)
p_load(tidyverse,DT,patchwork)

mpg %>% datatable()

mpg %>% 
  ggplot(aes(displ,hwy)) +
  geom_point()->p1

mpg %>% 
  ggplot(aes(displ,hwy)) +
  geom_point() +
  scale_x_continuous(breaks = seq(1,7,1),limits = c(1,7),expand = c(0,0)) +
  scale_y_continuous(breaks = seq(10,45,5),limits = c(10,45),expand = c(0,0))->p2

p1 + p2

9.1.1.1 数值格式

scales包提供了许多用于格式化数字标签的函数。其中最有用的是

• dollar
• comma
• percent

让我们用一些合成数据来演示这些函数。

# create some data
set.seed(1234)
df <- data.frame(xaxis = rnorm(50, 100000, 50000),
                 yaxis = runif(50, 0, 1),
                 pointsize = rnorm(50, 1000, 1000))
library(ggplot2)

df %>% datatable()

df %>% 
  ggplot(aes(xaxis,yaxis,size = pointsize)) +
  geom_point(color = "steelblue",
             alpha = .6) +
  scale_x_continuous(labels = scales::comma) +
  scale_y_continuous(labels = scales::percent) +
  scale_size_continuous(range = c(0,10),labels = scales::dollar)

若要将货币值格式化为欧元，可以使用

• label = scales::dollar_format(prefix = "“, suffix =”

df %>%
  ggplot(aes(xaxis, yaxis, size = pointsize)) +
  geom_point(color = "steelblue",
             alpha = .6) +
  scale_x_continuous(labels = scales::comma) +
  scale_y_continuous(labels = scales::percent) +
  scale_size_continuous(
    range = c(0, 10),
    label = scales::dollar_format(prefix = "", suffix = "\u20ac")
  )

9.1.2 分类数据轴

使用scale_x_discrete或scale_y_discrete函数修改分类轴。

选项包括

• limits-一个字符向量(定量变量按所需顺序的水平)
• labels-标签的字符向量(这些级别的可选标签)

mpg %>%
  ggplot(aes(class)) +
  geom_bar(fill = "steelblue") +
  scale_x_discrete(
    limits = c(
      "pickup",
      "suv",
      "minivan",
      "midsize",
      "compact",
      "subcompact",
      "2seater"
    ),
    labels = c(
      "Pickup\nTruck",
      "Sport Utility\nVehicle",
      "Minivan",
      "Mid-size",
      "Compact",
      "Subcompact",
      "2-Seater"
    )
  ) -> p1

mpg %>%
  count(class) %>%
  ggplot(aes(x = reorder(class, n), y = n)) +
  geom_col(fill = "steelblue") +
  scale_x_discrete(
    limits = c(
      "2seater",
      "minivan",
      "pickup",
      "subcompact",
      "midsize",
      "compact",
      "suv"
    ),
    labels = c(
      "2-Seater",
      "Minivan",
      "Pickup\nTruck",
      "Subcompact",
      "Mid-size",
      "Compact",
      "Sport Utility\nVehicle"
    )
  ) +
  labs(x = "class")-> p2

p1 / p2

9.1.3 时间轴

使用scale_x_date或scale_y_date函数修改日期轴。

选项包括：

• date_breaks -表示间隔时间的字符串，如“2周”或“10年”
• date_labels -给出标签格式规范的字符串

下表给出了日期值的格式规范。

Symbol	Meaning	Example
%d	day as a number	(0-31) 01-31
%a	abbreviated weekday	Mon
%A	unabbreviated weekday	Monday
%m	month (00-12)	00-12
%b	abbreviated month	Jan
%B	unabbreviated month	January
%y	2-digit year	07
%Y	4-digit year	2007

economics %>% 
  ggplot(aes(date,unemploy)) +
  geom_point(col = "red") +
  geom_line(col = "steelblue",size = 1) +
  scale_x_date(date_breaks = "5 years",date_labels = "%b-%y")

---

9.2 颜色

ggplot2图形中的默认颜色是功能性的，但通常在视觉上不具有足够的吸引力。幸运的是，这很容易改变。

特定的颜色可以是为点、线、条、区域和文本指定，或映射到数据集中变量的级别。

9.2.1 手动改变颜色

要为点、线或文本指定颜色，请在适当的geom中使用color = "colorname"选项。要为条形图和区域指定颜色，请使用fill = "colorname"选项。

• geom_point(color = “blue”)
• geom_bar(fill = “steelblue”)

要为变量的级别分配颜色，可以使用scale_color_manual和scale_fill_manual函数。前者用于指定点和线的颜色，而后者用于条形图和区域。

下面是一个示例，使用ggplot2附带的diamonds数据集。数据集包含了54,000颗圆形切割钻石的价格和属性。

diamonds$clarity %>% fct_count()

## # A tibble: 8 x 2
##   f         n
##   <fct> <int>
## 1 I1      741
## 2 SI2    9194
## 3 SI1   13065
## 4 VS2   12258
## 5 VS1    8171
## 6 VVS2   5066
## 7 VVS1   3655
## 8 IF     1790

## 手动指定颜色级别
diamonds %>% 
  ggplot(aes(cut,fill = clarity)) +
  geom_bar(position = "stack") +
  scale_fill_manual(values = c("darkred", "steelblue", 
                               "darkgreen", "gold",
                               "brown", "purple", 
                               "grey", "khaki4"))

diamonds %>% 
  ggplot(aes(cut,fill = clarity)) +
  geom_bar(position = "fill") +
  scale_fill_manual(values = c("darkred", "steelblue", 
                               "darkgreen", "gold",
                               "brown", "purple", 
                               "grey", "khaki4"))

diamonds %>% 
  ggplot(aes(cut,fill = clarity)) +
  geom_bar(position = "dodge") +
  scale_fill_manual(values = c("darkred", "steelblue", 
                               "darkgreen", "gold",
                               "brown", "purple", 
                               "grey", "khaki4"))

如果你和我一样在审美上有困难，另一种选择是使用预定义的调色板。

9.2.2 颜色调色板

在R中有许多预定义的调色板可用。

9.2.2.1 RColorBrewer

最流行的调色板可能是ColorBrewer调色板。

library(RColorBrewer)
RColorBrewer::display.brewer.all()

您可以使用scale_color_brewer和scale_fill_brewer函数指定这些调色板。

diamonds %>% 
  ggplot(aes(cut,fill = clarity)) +
  geom_bar() +
  scale_fill_brewer(palette = "Dark2")

diamonds %>% 
  ggplot(aes(cut,fill = clarity)) +
  geom_bar() +
  scale_fill_brewer(palette = "Dark2") +
  scale_x_discrete(expand = c(0,0)) +
  scale_y_continuous(expand = c(0,0))

向这些函数添加direction = -1将翻转调色板中颜色的顺序。

9.2.2.2 Viridis

viridis是另一个流行的调色板。

连续scales：

• scale_fill_viridis_c
• scale_color_viridis_c

离散scales:

• scale_fill_viridis_d
• scale_color_viridis_d

library(viridis)

## Loading required package: viridisLite

ggplot(diamonds,aes(x = cut,fill = clarity)) +
  geom_bar() +
  scale_fill_viridis_d()

9.2.2.3 其他调色板

其他需要探索的调色板包括dutchmasters、ggpomological、LaCroixColoR、nord、ochRe、palettetown、pals、 rcartocolor和wesanderson。如果您想查看所有选项板选项(或几乎所有选项)，请查看paletter包。

要了解有关颜色规范的更多信息，请参阅有关ggplot2颜色的R Cookpage页面。也可以参考这本书中的颜色选择建议这部分内容。

---

9.3 点线

9.3.1 Points

对于ggplot2图，默认的点是一个填充的圆。要指定不同的形状，请使用geom_point函数中的shape = #选项。要将形状映射到分类变量的级别，请使用aes函数中的shape = variablename选项。

例如：

• geom_point(shape = 1)
• geom_point(aes(shape = sex))

可用的形状如下表所示。

knitr::include_graphics("shapes1-1.png")

形状21到26提供了填充色和边框色。

9.3.2 Lines

默认的行类型是实线。要更改linetype，请使用geom_line函数中的linetype = #选项。要将linetype映射到类别变量的级别，请使用aes函数中的linetype = variablename选项。

例如：

• geom_line(linetype = 1)
• geom_line(aes(linetype = sex))

knitr::include_graphics("Lines2-1.png")

---

9.4 字体

R不太支持字体，但是通过一些工作，您可以更改图形中出现的字体。首先，您需要安装和设置extrafont包。

library(extrafont)

## Registering fonts with R

# extrafont::font_import()

# see what fonts are now available
fonts()

##   [1] "Agency FB"                       "Algerian"                       
##   [3] "Arial Black"                     "Arial"                          
##   [5] "Arial Narrow"                    "Arial Rounded MT Bold"          
##   [7] "Arial Unicode MS"                "Arvo"                           
##   [9] "Bahnschrift"                     "Baskerville Old Face"           
##  [11] "Bauhaus 93"                      "Bell MT"                        
##  [13] "Berlin Sans FB"                  "Berlin Sans FB Demi"            
##  [15] "Bernard MT Condensed"            "Blackadder ITC"                 
##  [17] "Bodoni MT"                       "Bodoni MT Black"                
##  [19] "Bodoni MT Condensed"             "Bodoni MT Poster Compressed"    
##  [21] "Book Antiqua"                    "Bookman Old Style"              
##  [23] "Bookshelf Symbol 7"              "Bradley Hand ITC"               
##  [25] "Britannic Bold"                  "Broadway"                       
##  [27] "Brush Script MT"                 "Calibri"                        
##  [29] "Calibri Light"                   "Californian FB"                 
##  [31] "Calisto MT"                      "Cambria"                        
##  [33] "Candara"                         "Candara Light"                  
##  [35] "Castellar"                       "Centaur"                        
##  [37] "Century"                         "Century Gothic"                 
##  [39] "Century Schoolbook"              "Chiller"                        
##  [41] "Colonna MT"                      "Comic Sans MS"                  
##  [43] "Consolas"                        "Constantia"                     
##  [45] "Cooper Black"                    "Copperplate Gothic Bold"        
##  [47] "Copperplate Gothic Light"        "Corbel"                         
##  [49] "Corbel Light"                    "Courier New"                    
##  [51] "Curlz MT"                        "DejaVu Sans Mono"               
##  [53] "DengXian"                        "DengXian Light"                 
##  [55] "Droid Serif"                     "Dubai"                          
##  [57] "Dubai Light"                     "Dubai Medium"                   
##  [59] "Ebrima"                          "Edwardian Script ITC"           
##  [61] "Elephant"                        "Engravers MT"                   
##  [63] "Eras Bold ITC"                   "Eras Demi ITC"                  
##  [65] "Eras Light ITC"                  "Eras Medium ITC"                
##  [67] "Euclid"                          "Euclid Extra"                   
##  [69] "Euclid Fraktur"                  "Euclid Math One"                
##  [71] "Euclid Math Two"                 "Euclid Symbol"                  
##  [73] "FangSong"                        "Felix Titling"                  
##  [75] "Fences"                          "Footlight MT Light"             
##  [77] "Forte"                           "Franklin Gothic Book"           
##  [79] "Franklin Gothic Demi"            "Franklin Gothic Demi Cond"      
##  [81] "Franklin Gothic Heavy"           "Franklin Gothic Medium"         
##  [83] "Franklin Gothic Medium Cond"     "Freestyle Script"               
##  [85] "French Script MT"                "FZShuTi"                        
##  [87] "FZYaoTi"                         "Gabriola"                       
##  [89] "Gadugi"                          "Garamond"                       
##  [91] "Georgia"                         "Gigi"                           
##  [93] "Gill Sans Ultra Bold"            "Gill Sans Ultra Bold Condensed" 
##  [95] "Gill Sans MT"                    "Gill Sans MT Condensed"         
##  [97] "Gill Sans MT Ext Condensed Bold" "Gloucester MT Extra Condensed"  
##  [99] "Goudy Old Style"                 "Goudy Stout"                    
## [101] "Haettenschweiler"                "Harlow Solid Italic"            
## [103] "Harrington"                      "High Tower Text"                
## [105] "HoloLens MDL2 Assets"            "Impact"                         
## [107] "Imprint MT Shadow"               "Indie Flower"                   
## [109] "Informal Roman"                  "Ink Free"                       
## [111] "Javanese Text"                   "Jokerman"                       
## [113] "Juice ITC"                       "KaiTi"                          
## [115] "Kristen ITC"                     "Kunstler Script"                
## [117] "Wide Latin"                      "Leelawadee"                     
## [119] "Leelawadee UI"                   "Leelawadee UI Semilight"        
## [121] "LiSu"                            "Lobster"                        
## [123] "Lucida Bright"                   "Lucida Calligraphy"             
## [125] "Lucida Console"                  "Lucida Fax"                     
## [127] "Lucida Handwriting"              "Lucida Sans"                    
## [129] "Lucida Sans Typewriter"          "Lucida Sans Unicode"            
## [131] "Magneto"                         "Maiandra GD"                    
## [133] "Malgun Gothic"                   "Malgun Gothic Semilight"        
## [135] "Marlett"                         "Matura MT Script Capitals"      
## [137] "Microsoft Himalaya"              "Microsoft Yi Baiti"             
## [139] "Microsoft New Tai Lue"           "Microsoft PhagsPa"              
## [141] "Microsoft Sans Serif"            "Microsoft Tai Le"               
## [143] "Microsoft Uighur"                "Microsoft YaHei"                
## [145] "Mistral"                         "Modern No. 20"                  
## [147] "Mongolian Baiti"                 "Monotype Corsiva"               
## [149] "MS Outlook"                      "MS Reference Sans Serif"        
## [151] "MS Reference Specialty"          "MT Extra"                       
## [153] "MT Extra Tiger"                  "MV Boli"                        
## [155] "Myanmar Text"                    "Niagara Engraved"               
## [157] "Niagara Solid"                   "Nirmala UI"                     
## [159] "Nirmala UI Semilight"            "OCR A Extended"                 
## [161] "Old English Text MT"             "Onyx"                           
## [163] "Open Sans"                       "Palace Script MT"               
## [165] "Palatino Linotype"               "Papyrus"                        
## [167] "Parchment"                       "Perpetua"                       
## [169] "Perpetua Titling MT"             "Playbill"                       
## [171] "Poiret One"                      "Poor Richard"                   
## [173] "Pristina"                        "Rage Italic"                    
## [175] "Raleway"                         "Ravie"                          
## [177] "Roboto"                          "Roboto Condensed"               
## [179] "Roboto Slab"                     "Rockwell"                       
## [181] "Rockwell Condensed"              "Rockwell Extra Bold"            
## [183] "Script MT Bold"                  "Segoe MDL2 Assets"              
## [185] "Segoe Print"                     "Segoe Script"                   
## [187] "Segoe UI"                        "Segoe UI Light"                 
## [189] "Segoe UI Semibold"               "Segoe UI Semilight"             
## [191] "Segoe UI Black"                  "Segoe UI Emoji"                 
## [193] "Segoe UI Historic"               "Segoe UI Symbol"                
## [195] "Showcard Gothic"                 "SimHei"                         
## [197] "SimSun-ExtB"                     "Snap ITC"                       
## [199] "STCaiyun"                        "Stencil"                        
## [201] "STFangsong"                      "STHupo"                         
## [203] "STKaiti"                         "STLiti"                         
## [205] "STSong"                          "STXihei"                        
## [207] "STXingkai"                       "STXinwei"                       
## [209] "STZhongsong"                     "Sylfaen"                        
## [211] "Symbol"                          "Symbol Tiger"                   
## [213] "Symbol Tiger Expert"             "Tahoma"                         
## [215] "Tempus Sans ITC"                 "Tiger"                          
## [217] "Tiger Expert"                    "Times New Roman"                
## [219] "Trebuchet MS"                    "Tw Cen MT"                      
## [221] "Tw Cen MT Condensed"             "Tw Cen MT Condensed Extra Bold" 
## [223] "Verdana"                         "Viner Hand ITC"                 
## [225] "Vivaldi"                         "Vladimir Script"                
## [227] "Webdings"                        "Wingdings"                      
## [229] "Wingdings 2"                     "Wingdings 3"                    
## [231] "YouYuan"                         "ZWAdobeF"

使用themes主题函数中的文本选项应用新字体。

library(extrafont)
ggplot(mpg, aes(x = displ, y=hwy)) +
  geom_point() +
  labs(title = "Diplacement by Highway Mileage",
       subtitle = "MPG dataset") +
  theme(text = element_text(size = 13, family = "Times New Roman"),
        plot.title = element_text(hjust = 0.5))

---

9.5 Legend

在ggplot2中，当变量被映射到颜色、填充、线型、形状、大小或alpha时，将自动创建图例。你对这些Legend的外观和感觉有很大的控制权。通常通过主题函数和/或labs函数进行修改。这里有一些最受欢迎的。

9.5.1 Legend位置

图例可以出现在图中的任何地方。默认情况下，它位于右侧。您可以使用：

theme(legend.position = position):

• “top” above the plot area
• “right” right of the plot area
• “bottom” below the plot area
• “left” left of the plot area c(x, y) within the plot area. The x and y values must range between 0 and 1. c(0,0) represents (left, bottom) and c(1,1) represents (right, top).
• “none” suppress the legend

例如，要将图例置于顶部，请使用以下代码。

mpg %>% 
  ggplot(aes(displ,hwy,col = class)) +
  geom_point(size = 4) +
  labs(title = "Diplacement by Highway Mileage") + 
  theme(legend.position = "top",
        legend.title.align = 0.5)  # Legend title align

9.5.2 Legend标题

您可以通过labs函数更改图例标题。使用颜色、填充、大小、形状、线条类型和alpha来给相应的图例赋予新的标题。图例标题的对齐是通过legend.title.align来控制的。对齐选项在主题函数。(0 =左,0.5 =中心,1 = 右)

# change the default legend title
ggplot(mpg, 
       aes(x = displ, y=hwy, color = class)) +
  geom_point(size = 4) +
  labs(title = "Diplacement by Highway Mileage",
       color = "Automobile\nClass") + 
  theme_minimal() +
  theme(legend.title.align=0.5)

mpg %>% 
  ggplot(aes(displ,hwy,col = class)) +
  geom_point(size = 4) +
  scale_color_discrete(name = "Automobile\nClass") +
  theme(legend.title.align = 0.5) +
  labs(title = "Diplacement by Highway Mileage")

---

9.6 Labels

标签是使图形易于理解的关键因素,可以使用labs函数添加。可用的选项如下所示

• title- main title
• subtitle- subtitle
• caption- caption (bottom right by default)
• x- horizontal axis
• y- vertical axis
• color- color legend title
• fill- fill legend title
• size- size legend title
• linetype- linetype legend title
• shape- shape legend title
• alpha- transparency legend title
• size- size legend title

ggplot(mpg, 
       aes(x = displ,
           y=hwy, 
           color = class,
           shape = factor(year))) +
  geom_point(size = 3, 
             alpha = .5) +
  labs(title = "Mileage by engine displacement",
       subtitle = "Data from 1999 and 2008",
       caption = "Source: EPA (http://fueleconomy.gov)",
       x = "Engine displacement (litres)",
       y = "Highway miles per gallon",
       color = "Car Class",
       shape = "Year") +
  theme(plot.title = element_text(hjust = 0.5),
        text = element_text(family = "Times New Roman"))

这不是一个很好的图——它太忙了，使得模式的识别非常困难。最好是将year变量分面。趋势线也会有帮助。

ggplot(mpg, 
       aes(x =displ, 
           y=hwy)) +
  geom_point(aes(col = class),
             size = 3, 
             alpha = .5) +
  geom_smooth(aes(col = class),method = "lm",se = FALSE) +
  labs(title = "Mileage by engine displacement",
       subtitle = "Data from 1999 and 2008",
       caption = "Source: EPA (http://fueleconomy.gov)",
       x = "Engine displacement (litres)",
       y = "Highway miles per gallon",
       color = "Car Class",
       shape = "Year") +
  theme(plot.title = element_text(hjust = 0.5),
        text = element_text(family = "Times New Roman")) +
  facet_wrap(~factor(year))

## `geom_smooth()` using formula 'y ~ x'

---

9.7 Annotations

注释是添加到图中的信息，用于突出重要的点。

9.7.1 添加文本

向图中添加文本有两个主要原因。一个是识别geom的数字质量。例如，我们可能希望在散点图中使用标签来标识点，或者在柱状图中标记条形图的高度。另一个原因是提供额外的信息，我们可能想要添加关于数据的注释，指出异常值等等。

9.7.1.1 Labling值

考虑以下散点图，基于mtcars数据集中的car数据。

mtcars %>% 
  ggplot(aes(wt,mpg)) +
  geom_point(size = 2)

让我们用它所代表的汽车的名称来标记每个点。

ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point(size = 2) +
  geom_text(label = row.names(mtcars))  # 说实话，不好看

# scatterplot with non-overlapping labels
data(mtcars)
library(ggrepel)
ggplot(mtcars, aes(x = wt,
                   y = mpg)) +
  geom_point(size = 2) +
  geom_text_repel(label = row.names(mtcars))

ggplot(mtcars, aes(x = wt, 
                   y = mpg)) +
  geom_point(size = 2) +
  geom_label_repel(label = row.names(mtcars))

9.7.1.2 添加附加信息

我们可以使用注释函数将文本放在图表的任何位置

annotate("text", 
         x, y, 
         label = "Some text", 
         color = "colorname", 
         size=textsize)

其中x和y是放置文本的坐标。颜色和大小参数是可选的。默认情况下，文本将居中。使用hjust和vjust来更改对齐方式。

• hjust 0 = left justified, 0.5 = centered, and 1 = right centered.
• vjust 0 = above, 0.5 = centered, and 1 = below.

继续前边的例子:

data(mtcars)
library(ggrepel)
txt <- paste("The relationship between car weight",
              "and mileage appears to be roughly linear",
              sep = "\n")
txt

## [1] "The relationship between car weight\nand mileage appears to be roughly linear"

ggplot(mtcars, aes(x = wt, 
                   y = mpg)) +
  geom_point(color = "red",
             size = 3) +
  geom_label_repel(label = row.names(mtcars), 
                   size=3) +
  ggplot2::annotate("text", 
                    6, 30, 
                    label=txt,
                    color = "red",
                    hjust = 1)

9.7.2 添加水平和垂直线

水平和垂直的线可以添加使用:

其中a是y轴上的数字，b是x轴上的数字。其他选项包括线型和颜色。

• geom_hline(yintercept = a)
• geom_vline(xintercept = b)

# add annotation line and text label
min_cty <- min(mpg$cty)
mean_hwy <- mean(mpg$hwy)
mean_cty <- mean(mpg$cty)
# 先绘制简单图，再逐渐修改细节
ggplot(mpg, 
       aes(x = cty, 
           y=hwy, 
           color=drv)) +
  geom_point(size = 3)+
  scale_y_continuous(limits = c(10,45,5),
                     breaks = seq(10,45,5)) +
  labs(color = "Drv") +
  scale_color_discrete(labels = c("4","F","R")) +
  theme(legend.position = "top") +
  geom_hline(yintercept = mean_hwy,
             color = "darkred",
             linetype = "dashed") +
  ggplot2::annotate("text", 
           min_cty, 
           mean_hwy + 1.5, 
           label = "Mean",
           color = "darkred") +
  geom_vline(xintercept = mean_cty,
             color = "darkred",
             linetype = "dashed") +
  ggplot2::annotate("text", 
           mean_cty + 1.5, 
           max(mpg$hwy), 
           label = "Mean",
           color = "darkred") +
  labs(title = "Mileage by drive type",
       x = "City miles per gallon",
       y = "Highway miles per gallon",
       color = "Drive") +
  theme(plot.title = element_text(hjust = 0.5))

我们还加上一条每加仑平均城市英里数的垂直线。在任何情况下，总是以某种方式标记注释行。否则读者将不知道他们的意思。

9.7.3 高亮某一组

有时您希望在图中突出显示单个组。gghighlight包中的gghighlight函数就是为此而设计的。

# highlight a set of points
library(ggplot2)
library(gghighlight)
ggplot(mpg, aes(x = cty, y = hwy)) +
  geom_point(color = "red",
             size=3) +
  gghighlight(class == "midsize")

mpg %>% 
  ggplot(aes(x = class)) +
  geom_bar()
# highlight a single bar
library(gghighlight)
ggplot(mpg, aes(x = class)) +
  geom_bar(fill = "red") +
  gghighlight(class == "midsize")

## Warning: Tried to calculate with group_by(), but the calculation failed.
## Falling back to ungrouped filter operation...

## label_key: class

这里没有什么是不能用基本图形来完成的，但是ggplot2更方便。

---

9.8 主题

ggplot2主题控制情节中所有与数据无关的组件的外观。您可以通过更改主题的元素来更改图形的外观和感觉。

9.8.1 改变主题元素

主题函数用于修改主题的各个组件。主题函数的参数在由在线帮助开发的备忘单中进行了描述。

考虑下图。它显示了2008-2009年某所大学按等级rank和学科discipline划分的男女教员人数。数据来源于Salaries数据集。

# create graph
data(Salaries, package = "carData")
p <- ggplot(Salaries, 
            aes(x = rank, fill = sex)) +
  geom_bar() 
p

data(Salaries, package = "carData")
p <- ggplot(Salaries, aes(x = rank, fill = sex)) +
  geom_bar() +
  facet_wrap(~discipline) +
  labs(title = "Academic Rank by Gender and Discipline",
       x = "Rank",
       y = "Frequency",
       fill = "Gender")
p

让我们改变一下主题。

• 将标签文本从黑色更改为海军蓝
• 将面板背景颜色从灰色改为白色
• 为主要的y轴网格线添加纯灰色线
• 添加小y轴网格线的虚线灰线
• 消除x轴网格线
• 将长条背景颜色改为白色，并带有灰色边框

p +
  theme(text = element_text(color = "navy"), # 将标签文本从黑色更改为海军蓝
        panel.background = element_rect(fill = "white"), # 将面板背景颜色从灰色改为白色
        panel.grid.major.y = element_line(color = "grey"), # 为主要的y轴网格线添加纯灰色线
        panel.grid.minor.y = element_line(color = "grey", 
                                          linetype = "dashed"),
        panel.grid.major.x = element_blank(),
        panel.grid.minor.x = element_blank(), # 消除x轴网格线
        strip.background = element_rect(fill = "white", color="grey"))

哇，这看起来很糟糕，但你明白我的意思。

9.8.1.1 ggThemeAssist

如果您想使用GUI创建自己的主题，请查看ggThemeAssist。安装包后，RStudio中的Addins下将出现一个新菜单项。

突出显示创建图形的代码，然后从Addins下拉菜单中选择ggThemeAssist选项。您可以使用指向-单击来更改主题的许多特性。当您完成时，主题代码将被附加到您的图形代码中。

9.8.2 Pre-packaged主题

我不是一个很好的艺术家(只看最后一个例子)，所以我经常寻找可以应用到我的图表中的预先包装好的主题。有很多可用的主题,有些ggplot2附带。其中包括theme_classic、theme_dark、theme_gray、theme_grey、theme_light theme_linedraw、theme_minimum和theme_void。在本书中，我们经常使用theme_minimum。其他的可以通过附加包获得。

9.8.2.1 ggthemes

ggthemes package覆盖19种主题themes.

Theme	Description
theme_base	Theme Base
theme_calc	Theme Calc
theme_economist	ggplot color theme based on the Economist
theme_economist_white	ggplot color theme based on the Economist
theme_excel	ggplot color theme based on old Excel plots
theme_few	Theme based on Few’s “Practical Rules for Using Color in Charts”
theme_fivethirtyeight	Theme inspired by fivethirtyeight.com plots
theme_foundation	Foundation Theme
theme_gdocs	Theme with Google Docs Chart defaults
theme_hc	Highcharts JS theme
theme_igray	Inverse gray theme
theme_map	Clean theme for maps
theme_pander	A ggplot theme originated from the pander package
theme_par	Theme which takes its values from the current ‘base’ graphics parameter values in ‘par’.
theme_solarized	ggplot color themes based on the Solarized palette
theme_solarized_2	ggplot color themes based on the Solarized palette
theme_solid	Theme with nothing other than a background color
theme_stata	Themes based on Stata graph schemes
theme_tufte	Tufte Maximal Data, Minimal Ink Theme
theme_wsj Wall	Street Journal theme

为了演示它们的用法，我们首先创建并保存一个图。

# create basic plot
library(ggplot2)
p <- ggplot(mpg, 
            aes(x = displ, 
                y=hwy, 
                color = class)) +
  geom_point(size = 3, 
             alpha = .5) +
  labs(title = "Mileage by engine displacement",
       subtitle = "Data from 1999 and 2008",
       caption = "Source: EPA (http://fueleconomy.gov)",
       x = "Engine displacement (litres)",
       y = "Highway miles per gallon",
       color = "Car Class") +
  theme(plot.title = element_text(hjust = 0.5))

p
# display graph
p + theme(panel.grid.major = element_line(colour = "green", 
    linetype = "dotdash"), panel.grid.minor = element_line(linetype = "dotdash"), 
    panel.background = element_rect(fill = "gray84", 
        linetype = "dotdash"), plot.background = element_rect(fill = "antiquewhite", 
        colour = NA))

让我们应用一些主题

# add economist theme
library(ggthemes)
p + theme_economist() 

# add fivethirtyeight theme
p + theme_fivethirtyeight()

# add wsj theme
p + theme_wsj(base_size=8)

默认情况下，wsj主题的字体通常太大。更改base_size选项会有所帮助。

每个主题还提供了颜色和填充比例。在下一个示例中，将同时使用少数几个主题和颜色。

# add few theme
p + theme_few() + scale_color_few()

9.8.2.2 hrbrthemes

hrbrthemes包专注于以排版为中心的主题。结果是图表，往往有一个干净的外观。继续上面的示例图.

# add few theme
library(hrbrthemes)
p + theme_ipsum()

9.8.2.3 ggthemer

ggthemer包提供了广泛的主题(截至打印时为17个主题)。该软件包在CRAN上不可用，必须从GitHub上安装。

library(devtools)

## Loading required package: usethis

# install_github('cttobin/ggthemr')
p_load(ggthemr)

函数的工作方式略有不同。使用ggthemr(“themename”)函数将未来的图形设置为给定的主题。使用ggthemr_reset()将未来的图返回到ggplot2默认主题。

主题主要包括：flat, flat dark, camoflauge, chalk, copper, dust, earth, fresh, grape, grass, greyscale, light, lilac, pale, sea, sky, and solarized.

# set graphs to the flat dark theme
library(ggthemr)
ggthemr("flat dark")
p

我不会把这个主题用在这个图上。很难分辨颜色。哪个绿色代表紧凑型汽车，哪个代表微型车?选择一个能最好地将图表信息传达给受众的主题。

ggthemr_reset()
p
p + scale_color_brewer(palette = "Accent")