Figures

Quarto includes a number of features aimed at making it easier to work with figures and subfigures, as well as for laying out panels that contain multiple figures, tables, or other content.

Figure Basics

In Pandoc markdown, a figure is created whenever a captioned image appears by-itself in a paragraph. For example:

![Elephant](elephant.png)

This results in the following treatment for various output types:

HTML PDF Word
A line drawing of an elephant. A line drawing of an elephant. A line drawing of an elephant.

Note that for LaTeX / PDF output figures are automatically numbered (you can arrange for figures to be numbered in other formats using Cross References).

Linked Figures

When rendering with Quarto, you can enclose a figure within a link and it will still be treated within output as a captioned figure. For example:

[![Elephant](elephant.png)](https://en.wikipedia.org/wiki/Elephant)

Figure Alignment

Figures and their captions are center aligned by default. Add the fig-align attribute to the image to use a different alignment. For example:

![Elephant](elephant.png){fig-align="left"}

Alt Text

You can add alternative text to a figure by adding the fig-alt attribute to the image. For example, the following Markdown…

![](elephant.png){fig-alt="A drawing of an elephant."}

… will create the following HTML with the corresponding alt tag:

<img src="elephant.png" alt="A drawing of an elephant.">

Note that the figure caption, title, and alt text can all be different. For example, the following code…

![Elephant](elephant.png "Title: An elephant"){fig-alt="A drawing of an elephant."}

…produces this HTML:

<img src="elephant.png" title="Title: An elephant" alt="A drawing of an elephant.">

Multiformat Figures

You can write markdown that provides a distinct image file format depending on the target output format. To do this simply leave-off the extension, for example:

![](elephant)

By default this will look for elephant.png, however if you are rendering to PDF it will look for elephant.pdf. You can customize this behavior using the default-image-extension option. For example:

format:
  html:
    default-image-extension: svg
  pdf:
    default-image-extension: tex

Subfigures

If you have several figures that appear as a group, you can create a figure div to enclose them. For example:

::: {#fig-elephants layout-ncol=2}

![Surus](surus.png){#fig-surus}

![Hanno](hanno.png){#fig-hanno}

Famous Elephants
:::

Again, the last paragraph provides the main caption, and the individual figures carry the sub-captions. Here is what this looks like when rendered as HTML:

An artistic rendition of Surus, Hannibal's last war elephant, is on the left. Underneath this picture is the caption '(a) Surus.' On the right is a line drawing of Hanno, a famous elephant. Underneath this picture is the caption '(b) Hanno.' The words 'Figure 1: Famous elephants' are centered beneath both pictures.

Note that we also used a layout-ncol attribute to specify a two-column layout. The next section delves more into customizing figure layouts.

Figure Panels

Above we demonstrate laying out two side-by-side figures with subcaptions and a main caption. You may or may not want the caption / sub-caption treatment, and you might also want to use multiple rows of figures. All of these variations are possible.

To layout two figures with their own standalone captions (and no main caption), just eliminate the #fig identifiers and main caption at the bottom:

::: {layout-ncol=2}
![Surus](surus.png)

![Hanno](hanno.png)
:::

An artistic rendition of Surus, Hannibal's last war elephant, is on the left. Underneath this picture is the caption 'Surus.' On the right is a line drawing of Hanno, a famous elephant. Underneath this picture is the caption 'Hanno.'

You can also eliminate the captions entirely:

::: {layout-ncol=2}
![](surus.png)

![](hanno.png)
:::

Multiple Rows

If you have more than 2 images, you might want to lay them out across multiple rows. You can do this using the layout-nrow attribute. For example:

::: {layout-nrow=2}
![Surus](surus.png)

![Hanno](hanno.png)

![Abdul Abbas](abdul-abbas.png)

![Lin Wang](lin-wang.png)
:::

A 2x2 grid of pictures of elephants. There are labels underneath each of the pictures. Clockwise from the upper left, the labels say: Surus, Hanno, Lin Wang, and Abdul Abbas.

More complex figure arrangements (e.g. rows with varying column layouts) are possible. See the Custom Layouts section below for more details.

Figure Divs

You can treat any markdown content you want as a figure by enclosing it in Pandoc div block with an identifier prefaced with #fig-. For example, here we create a figure that includes an embedded iframe:

::: {#fig-elephant}

<iframe width="560" height="315" src="https://www.youtube.com/embed/SNggmeilXDQ"></iframe>

Elephant
:::

Note that the last paragraph in the div block is used as the figure caption.

LaTeX Figures

This section describes some figure handling options that are specific to LaTeX output.

One very important thing to note is that using the fig-env and fig-pos options described below will trigger the creation of a LaTeX floating environment (most often \begin{figure}). Depending upon where this LaTeX is generated it (e.g. is it within another \begin{figure}) this could generate invalid LaTeX. To be on the safe side these attributes should typically only be used for images at the very top level of your document.

Environments

There are a number of LaTeX packages that provide custom figure environments. For example, the mdframed package includes an mdframed environment used to enclose figures in a special border style. By default, Quarto uses the standard figure environment, but you can use the fig-env attribute to specify a custom one. For example:

---
title: "Sidenotes"
format:
  pdf:
    include-in-header: 
      text: |
        \usepackage{mdframed}
---

![Elephant](elephant.png){fig-env="mdframed"}

Figure Position

The default LaTeX figure is a floating environment, so the specific location it appears in your document will depend on its size and the nature of the other content around it. You can exercise some control over this behavior using the fig-pos option. The fig-pos option provides a placement specifier for the figure environment. For example, the ht in this LaTeX snippet is the fig-pos:

\begin{figure}[ht]

\end{figure}

You can specify fig-pos either at the document level, as a executable code block option, or within markdown. Here we do all three:

---
title: "My Document"
format:
  pdf:
    fig-pos: 'h'
---

```{python}
#| fig-pos: 't'

```

![](figure.png){fig.pos='b'}

See this article for additional details on LaTeX figure positioning.

Figures and Code Blocks

If your figure was generated by an executable code block and the code was included in the output (echo: true), then fig-pos will be set to H by default (to try to keep it alongside the code that generated it). In all other cases, default LaTeX handing of figure position is used unless you specify an explicit fig-pos.

PGF/TikZ Graphics

If you are creating LaTeX output, Quarto will automatically emit the correct LaTeX for markdown images that reference .tex files containg PGF/TikZ vector graphics. For example, the following markdown images:

![](image1.tex)

![](image2.tex){width=80%}

Will be written to LaTeX as:

\input{image1.tex}

\resizebox{0.8\linewidth}{!}{\input{image2.tex}}

As shown above, width and height percentages are automatically converted to \linewidth scaled. Alternatively you can specify custom LaTeX for the width and height arguments of \resizebox.

Caption Locations

By default, figure captions are positioned below figures. You can modify this behavior using the fig-cap-location option. For example:

---
fig-cap-location: top
---

Note that this option is specified at the top level so that it can be shared by both PDF and HTML formats. If you are only targeting a single format you can place it alongside other format specific options.

Valid values for the caption location include:

Value Description
top Position the caption above the figure.
bottom Position the caption below the figure.
margin Position the caption in the margin.

See the article on Article Layout for additional details on placing captions in the margin.

Custom Layouts

The examples above used the layout-ncol or layout-nrow attributes to create straightforward layouts where all columns are of equal sizes. The layout attribute enables the creation of much more complex layouts.

For example, this defines a layout with two equally sized figures in the first row, then another image that spans the entire second row:

::: {layout="[[1,1], [1]]"}
![Surus](surus.png)

![Hanno](hanno.png)

![Lin Wang](lin-wang.png)
:::

Three elephant pictures arranged such that two pictures are side-by-side in the first row, and the third picture is underneath both of these. The picture on the left in the first row is captioned 'Surus' and the picture on the right is captioned 'Hanno'. The picture underneath these two is captioned 'Lin Wang' and is as as wide and tall as the other two put together.

The layout attribute is a 2-dimensional array where the first dimension defines rows and the second columns. In this case "layout="[[1,1], [1]]" translates to: create two rows, the first of which has two columns of equal size and the second of which has a single column.

Note that the numbers in a row are arbitrary and don’t need to add up to a particular total. You can therefore use whatever scheme is most natural. For example, here we define columns that occupy varying percentage widths of the row:

::: {layout="[[70,30], [100]]"}
![Surus](surus.png)

![Hanno](hanno.png)

![Lin Wang](lin-wang.png)
:::

You can also use negative values to create space between elements. For example:

::: {layout="[[40,-20,40], [100]]"}
![Surus](surus.png)

![Hanno](hanno.png)

![Lin Wang](lin-wang.png)
:::

Three elephant pictures arranged such that two pictures are side-by-side in the first row, and the third picture is underneath both of these. The picture on the left in the first row is captioned 'Surus' and the picture on the right is captioned 'Hanno'. The two pictures are separated by some whitespace. The picture underneath these two is captioned 'Lin Wang' and is wider and taller than the other two put together.

Vertical Alignment

If you have a layout with a row of images of differing heights, you can control their vertical alignment using the layout-valign attribute. A simple example:

::: {layout="[25,-2,10]" layout-valign="bottom"}
![Surus](surus.png)

![Lin Wang](lin-wang.png)
:::

Two pictures of elephants side by side. The picture on the left has the label 'Figure 1: Surus' underneath it. The picture on the right has the label 'Figure 2: Lin Wang' underneath it. The figure on the left is more than twice the width and height of the figure on the right.

Note that vertical alignment isn’t limited to images, you can also vertically align any other elements that are included in a panel.

Computations

Figures

Note that figure layout attributes also work for figures produced by executable code blocks. Here are examples for both Jupyter and Knitr:

```{python}
#| layout-ncol: 2

import matplotlib.pyplot as plt
plt.plot([1,23,2,4])
plt.show()

plt.plot([8,65,23,90])
plt.show()
```

Two line plots rendered by Jupyter side-by-side.

```{r}
#| layout-ncol: 2

plot(cars)
plot(pressure)
```

Two scatter plots arranged side-by-side.

Figure Captions

You can also include figure and subfigure captions on computational output using the fig-cap and fig-subcap options. When applying captions to computational output you can optionally include a label with a fig- prefix—if you do this then the figure will be numbered and cross-referenceable.

```{python}
#| label: fig-charts
#| fig-cap: Charts
#| fig-subcap: ["First", "Second"]
#| layout-ncol: 2

import matplotlib.pyplot as plt
plt.plot([1,23,2,4])
plt.show()

plt.plot([8,65,23,90])
plt.show()
```

Two line plots with captions rendered by Jupyter side-by-side.

```{r}
#| label: fig-charts
#| fig-cap: Charts
#| fig-subcap: ["Cars", "Pressure"]
#| layout-ncol: 2

plot(cars)
plot(pressure)
```

Two scatter plots with captions arranged side-by-side.

Custom Layout

The layout works the same way for figures produced by Knitr or Jupyter. For example, here’s an Rmd code chunk that produces 3 plots and defines a custom layout for them:

```{r}
#| layout: [[45,-10, 45], [100]]

plot(cars)
plot(pressure)
plot(mtcars)
```

Two plots arranged side-by-side with a large plot underneath it. The top two plots are scatter plots visualizing the `cars` and `pressure` datasets. These two plots are separated by some additional white space. The plot on the bottom visualizes the `mtcars` dataset and is wider and taller than the other two plots combined. This plot is an 11 by 11 grid plotting each of the 11 variables in the `mtcars` dataset against each other as a scatterplot. Instead of scatter plots in the diagonal starting in the upper left and going to the lower right are text labels for each of the variable names. These are: 'mpg', 'cyl', 'disp', 'hp', 'drat', 'wt', 'qsec', 'vs', 'am', 'gear', and 'carb'.

Block Layout

While the examples above illustrate laying out figures, it’s important to note that layout attributes can be used to layout any sort of block content. For example, here we layout 2 lists side-by-side:

::: {layout-ncol=2}
### List One

- Item A
- Item B
- Item C

### List Two

- Item X
- Item Y
- Item Z
:::

Note that headings are automatically combined with the block that follows them, so this markdown has a total of 2 columns to lay out. Here’s an example of a paragraph next to a bullet list (without headings):

::: {layout-ncol=2}
- Item X
- Item Y
- Item Z

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Curabitur gravida eu erat et fring. Morbi congue augue vel eros ullamcorper, eget convallis tortor sagittis. Fusce sodales viverra mauris a fringilla. Donec feugiat, justo eu blandit placerat, enim dui volutpat turpis, eu dictum lectus urna eu urna. Mauris sed massa ornare, interdum ipsum a, semper massa. 
:::