r - multi-panel scatterplot in Cars package

Question

I'm trying to make a 2 panel figure in RStudio. Simple enough usually:

par(mfrow=c(1,2)) #1*2 plotting window

However, when I make the plot using the scatterplot() function from the car package it seems to override this division of the plotting window.

Reproducible example: The format I'm looking for can be produced with:

par(mfrow=c(1,2))
plot(iris$Sepal.Length,iris$Sepal.Width)
plot(iris$Petal.Width,iris$Sepal.Width)

But I'd like to use scatterplot() for a few reasons. But when I try the same formatting trick again, it doesn't work. Try this block:

library(car)
par(mfrow=c(1,2))
scatterplot(Sepal.Length~Sepal.Width|Species,data=iris,grid="FALSE", boxplots="", reg.line="FALSE",pch=c(0,1,2))
scatterplot(Petal.Width~Sepal.Width|Species,data=iris,grid="FALSE", boxplots="", reg.line="FALSE",pch=c(0,1,2))

Is anyone aware of a workaround for this?

I also tried an alternative structuring with:

m<-rbind(c(1,2))
layout(m)

No luck.

UPDATE This question is largely a duplicate of this one, except I'm looking for a workaround for it rather than the responses left on that page which basically just acknowledge that scatterplot() does in fact override par(mfrow)

Answer 1

rmarkdown您可以使用 chunk 参数在文档中并排输出数字fig.show='hold'。另一种选择是用于ggplot2创建绘图，而不是cars scatterplot函数。我在下面展示了这两种方法。

文档中的并排cars::scatterplot图rmarkdown

以下是 PDF 输出的示例。设置输出文档中每个绘图的实际大小，与和out.width='3in'无关。但是您仍然可以调整和调整相对于绘图区域的纵横比和文本大小。fig.heightfig.widthfig.heightfig.width

---
title: "Untitled"
author: "eipi10"
date: "November 23, 2016"
output: pdf_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r fig.show='hold', out.width='3in', fig.height=5, fig.width=5}
library(car)

scatterplot(Sepal.Length~Sepal.Width|Species,data=iris,grid="FALSE",
            boxplots="", reg.line="FALSE", pch=c(0,1,2))
scatterplot(Petal.Width~Sepal.Width|Species,data=iris,grid="FALSE", 
            boxplots="", reg.line="FALSE", pch=c(0,1,2))
```

并排散点图使用ggplot2

如果您愿意，ggplot2那么您可以相对轻松地获得两个情节布局：

library(ggplot2)
library(gridExtra)
theme_set(theme_bw())

grid.arrange(
  ggplot(iris, aes(Sepal.Width, Sepal.Length, colour=Species)) +
    geom_point() +
    geom_smooth(alpha=0.2) +
    theme(legend.position="top"),
  ggplot(iris, aes(Sepal.Width, Petal.Width, colour=Species)) +
    geom_point() +
    geom_smooth(alpha=0.2) +
    theme(legend.position="top"),
  ncol=2)

自定义ggplot2绘图以看起来更像cars::scatterplot输出

您可以通过其他方式自定义上面的代码。如果您不想要置信区间，请添加se=FALSE到geom_smooth. 如果您希望每个物种有不同的形状，请添加aes(shape=Species)到geom_point. 如果您想要基本图形中使用的特定形状，请添加+ scale_shape_manual(values=0:2)等。您还可以通过一些额外的工作获得单个图例。

在下面的代码中，我添加了这些和其他自定义项，以重现更接近原始基本图形的内容。

# Components we'll reuse for both plots
my_theme = list(geom_point(aes(shape=Species)),
                geom_smooth(se=FALSE, show.legend=FALSE, lwd=0.8),
                scale_shape_manual(values=0:2),
                scale_colour_manual(values=c("black", "red","green")),
                theme_bw(),
                theme(panel.grid.major=element_blank(),
                      panel.grid.minor=element_blank(),
                      legend.position="top"))

p1 = ggplot(iris, aes(Sepal.Width, Sepal.Length, colour=Species)) +
  my_theme +
  labs(x="Sepal Width", y="Sepal Length") +
  scale_y_continuous(limits=c(3,8)) +
  scale_x_continuous(limits=c(1,5)) 

p2 = ggplot(iris, aes(Sepal.Width, Petal.Width, colour=Species)) +
  my_theme +
  labs(x="Sepal Width", y="Petal Width")

# Function to extract legend
# https://github.com/hadley/ggplot2/wiki/Share-a-legend-between-two-ggplot2-graphs
g_legend<-function(a.gplot){
  tmp <- ggplot_gtable(ggplot_build(a.gplot))
  leg <- which(sapply(tmp$grobs, function(x) x$name) == "guide-box")
  legend <- tmp$grobs[[leg]]
  return(legend)}

leg = g_legend(p1)

grid.arrange(leg,
  arrangeGrob(grobs=lapply(list(p1,p2), function(p) p + guides(colour=FALSE, shape=FALSE)), ncol=2), 
  ncol=1, heights=c(1,10))