09-Arranging_plots.Rmd

# Arranging Plots

**Learning Objectives**

1. Produce several subplots as part of the same main visualization
2. Understand how to arrange plots 
    - in rows and columns
    - on top of each other (insets)



## Introduction {-}

Want to arrange multiple plots; need more than faceting

This chapter discusses {patchwork}. 

Others:

  * {cowplot}
  * {gridExtra}
  * {ggpubr}

## Arranging plots side by side with no overlap {-}

```{r 09-library, message=FALSE, warning=FALSE, include=FALSE, paged.print=FALSE}
library(tidyverse)
```

Let's imagine we have several plots:

```{r 09-dataplots}
p1 <- ggplot(mpg) + 
  geom_point(aes(x = displ, y = hwy))

p2 <- ggplot(mpg) + 
  geom_bar(aes(x = as.character(year), fill = drv), position = "dodge") + 
  labs(x = "year")

p3 <- ggplot(mpg) + 
  geom_density(aes(x = hwy, fill = drv), colour = NA) + 
  facet_grid(rows = vars(drv))

p4 <- ggplot(mpg) + 
  stat_summary(aes(x = drv, y = hwy, fill = drv), geom = "col", fun.data = mean_se) +
  stat_summary(aes(x = drv, y = hwy), geom = "errorbar", fun.data = mean_se, width = 0.5)
```


```{r 09-plot1}
library(patchwork)

p1 + p2 # same as p1 | p2
```

```{r 09-plot2}
p1 + p2 + p3 + p4
```


## Controlling the layout {-}

```{r 09-plot3}
p1 + p2 + p3 + plot_layout(ncol = 2)
```

```{r 09-plot4}
p1 / p2
```

## More compositions: {-}

```{r 09-plot5}
p3 | (p2 / (p1 | p4))
```

## Layouts can get creative! {-}

```{r 09-plot6}
layout <- "
AAB
C#B
CDD
"

p1 + p2 + p3 + p4 + plot_layout(design = layout)
```

## Collect repeats of the same legend {-}

```{r 09-plot7}
p1 + p2 + p3 + guide_area() + plot_layout(ncol = 2, guides = "collect")
```


## Parts of the patchwork object can still be modified {-}

```{r 09-plot8}
p12 <- p1 + p2
p12[[1]] <- p12[[1]] + theme_light()
p12
```

## New operator: `&` adds whole-plot themes {-}

```{r 09-plot9}
p1 + p4 & 
  scale_y_continuous(limits = c(0, 45)) & 
  theme_minimal()
```

### Plot annotations {-}
```{r 09-plot10}
p34 <- p3 + p4 + 
  plot_annotation(
  title = "A closer look at the effect of drive train in cars",
  caption = "Source: mpg dataset in ggplot2",
  tag_levels = "I") &
  theme_gray(base_family = "mono")

p34
```

## Labeling plots (e.g. parts of figures) {-}

```{r 09-plot11}
new <- p34
keep <- p34
new[[2]] <- new[[2]] + plot_layout(tag_level = "new")
keep[[2]] <- keep[[2]] + plot_layout(tag_level = "keep")
```

```{r}
new
```

```{r}
keep
```


## Specify the type of tags/labels {-}
```{r}
p34 + plot_annotation(tag_levels = c("I", "a"))
```


## Arranging plots on top of each other {-}

Instead of putting plots next to or above/below each other, we can nest them.

General options: left, right, top, and bottom.
Set specific locations

- `grid::unit()`--npc units ("Normalized Parent Coordinates"), ranges from 0 to 1

Location is relative to **panel area**, not **plot area**. `align_to` arg changes this.

## Example: {-}

Placing an inset exactly 15 mm from the top right corner:

```{r 09-plot12}
p1 + 
  inset_element(
    p2, 
    left = 0.4, 
    bottom = 0.4, 
    right = unit(1, "npc") - unit(15, "mm"), 
    top = unit(1, "npc") - unit(15, "mm"),
    align_to = "full"
  )
```

## Another inset example with annotation {-}

```{r 09-plot13}
p24 <- p2 / p4 + plot_layout(guides = "collect")

p1 + 
  inset_element(p24, left = 0.5, bottom = 0.05, right = 0.95, top = 0.9)  +
  plot_annotation(tag_levels = 'A') & 
  theme_bw()
```



## Extra {-}

**grid and gridExtra packages**
```{r 09-plot14, message=FALSE, warning=FALSE, paged.print=FALSE}
library(gridExtra)
library(grid)

grid.arrange(p1, p2, p3, p4, ncol=2)
```


**cowplot package**
```{r 09-plot15, message=FALSE, warning=FALSE, paged.print=FALSE}
library(cowplot)

plot_grid(p1, p2, 
          labels = c('A', 'B'),
          align = "h",
          label_size = 12,
          label_fontfamily = "serif",
          label_fontface = "plain",
          label_colour = "blue",
          rel_widths = c(1, 2))
```


```{r 09-plot16, message=FALSE, warning=FALSE, paged.print=FALSE}
# We saved this to the images folder already.

# ggsave("images/p1.png", p1)

ggdraw()+
  draw_plot(p3)+
  draw_image("images/p1.png",x=0.4,y=0.4,scale=0.5)+
  draw_label("This is some text",x=0.5,y=0.9)
```



To add a common title we use `ggdraw()
```{r 09-plot17}
plot_row <- plot_grid(p1, p2)

# now add the title
title <- ggdraw() + 
  draw_label("Common title", 
             fontface = 'bold', x = 0, hjust = 0) +
  theme(plot.margin = margin(0, 0, 0, 7))

plot_grid(title, plot_row,
          ncol = 1,
          rel_heights = c(0.1, 1))
```

```{r 09-plot18}
ggdraw()+
  draw_plot(p1,x=0.1,y=0,scale=0.5)+
  draw_plot(p2,x=-0.2,y=0,scale=0.5)+
  draw_label("This plot is made with cowplot",
             x=0.5,y=0.95,
             color="pink",
             fontface="bold",size=20)
```

**ggpubr package**
```{r 09-plot19, message=FALSE, warning=FALSE, paged.print=FALSE}
library(ggpubr)
ggarrange(p1, p2, p3, ncol = 2, nrow = 2,common.legend = TRUE)
```


## Conclusions {-}


[Patchwork - imaginist](https://patchwork.data-imaginist.com/) is one of the packages mentioned in the book, also some other packages provide same results with different approaches. 

### Extra resources: {-}

- [grid and gridExtra](https://cran.r-project.org/web/packages/gridExtra/vignettes/arrangeGrob.html)
- [cowplot](https://wilkelab.org/cowplot/articles/plot_grid.html)
- [ggpubr](https://rpkgs.datanovia.com/ggpubr/reference/ggarrange.html)


## Meeting Videos {-}

### Cohort 1 {-}

`r knitr::include_url("https://www.youtube.com/embed/mGE639JpneE")`

<details>
  <summary> Meeting chat log </summary>
```
00:27:47	Lydia Gibson:	What are npc unts?
00:27:56	Michael Haugen:	"npc" (Normalized Parent Coordinates)
00:28:02	Michael Haugen:	0 to 1
00:28:07	Lydia Gibson:	Oh okay. Thank you!
00:28:16	Michael Haugen:	Same thing that was used for faceting annotations.
00:28:43	Michael Haugen:	so .8 is l80 percent of the way up the y axis for example.
00:28:47	Lydia Gibson:	I missed annotations last week. I’ll have to go back and watch the session.
00:43:21	SriRam:	I use patch and cowplot
00:50:00	Kent Johnson:	https://www.cedricscherer.com/2019/08/05/a-ggplot2-tutorial-for-beautiful-plotting-in-r/
00:50:09	Lydia Gibson:	Thank you!
00:50:17	Michael Haugen:	The arrows are in chapter 8.3 with geom and curve for example, annotate(
    geom = "curve", x = 4, y = 35, xend = 2.65, yend = 27, 
    curvature = .3, arrow = arrow(length = unit(2, "mm"))
  ) +
00:51:16	Michael Haugen:	and arrows came up in the discussion as a discussion of the GROB and arrows and how to render your plot so the arrows are not distorted.
00:51:16	Ryan Metcalf:	Perfect! “arrow” was the argument I was after!
00:51:37	Michael Haugen:	And then we talked about ggsave as a part of that
00:53:32	Michael Haugen:	we all will be at Cedric’s level by the end of this bookclub right?
00:53:47	SriRam:	:D
00:53:52	Lydia Gibson:	Hopefully lol
00:54:20	Ryan S:	Thank you!
00:54:29	SriRam:	Thank you
```
</details>

`r knitr::include_url("https://www.youtube.com/embed/lxkk5MCWREQ")`

<details>
  <summary> Meeting chat log </summary>
```
00:04:54	June Choe:	https://yjunechoe.github.io/ggtrace-talk/
00:12:16	Ryan S:	brilliant... didn't know this before but really simplifies the concept
00:14:08	Michael Haugen:	Makes sense
00:26:15	June Choe:	https://ggplot2.tidyverse.org/reference/aes_eval.html
00:26:36	Ryan Metcalf:	I’m thinking in the context….I buy a car. The engineers have optimized it for longevity….but I want a hot rod….So I need to open the hood and change parts. Or, access the computer and start changing parameters.
00:32:54	SriRam:	This is like scuba diving, more beautiful under the surface :)
00:33:15	Stan Piotrowski:	Great analogy, SriRam!
00:33:17	Ryan Metcalf:	Completely agree @SriRam!
00:36:44	June Choe:	ggplot2:::ggplot_build.ggplot
00:37:58	June Choe:	ggplot2:::print.ggplot
00:50:22	Federica Gazzelloni:	thanks June!!!
00:50:54	SriRam:	Out of curiosity, how much of this trickery (internal functions) can be learnt from "advanced R" or are these mentioned in the ggplot book ? I am just a regular user, I may not go this deep, but looks very interesting to explore/read during the Christmas break
00:51:12	Stan Piotrowski:	I’m in the same boat as SriRam
00:51:32	Stan Piotrowski:	Curious to know more about this but can definitely see myself getting lost in a rabbit hole
00:54:23	Ryan S:	at some point -- maybe a different session -- can we dive deep into the different stat options ("identity", "count", etc.)
00:54:47	Ryan S:	specifically, what do they do and when would you use them
00:55:09	Ryan Metcalf:	June, this is amazing!
00:57:27	SriRam:	Countdown starts...... 5 mins to come back to reality !!! :D
01:02:21	Stan Piotrowski:	Great talk, June!
01:02:34	Kent Johnson:	Thank you! See you next week!
```
</details>