Insert picture/table in R Markdown
When it comes to inserting a picture, r2evans's suggestion of ![Caption for the picture.](/path/to/image.png)
can be problematic if PDF output is required.
The knitr function include_graphics
knitr::include_graphics('/path/to/image.png')
is a more portable alternative
that will generate, on your behalf, the markdown that is most appropriate to the output format that you are generating.
Update: since the answer from @r2evans, it is much easier to insert images into R Markdown and control the size of the image.
Images
The bookdown book does a great job of explaining that the best way to include images is by using include_graphics()
. For example, a full width image can be printed with a caption below:
```{r pressure, echo=FALSE, fig.cap="A caption", out.width = '100%'}
knitr::include_graphics("temp.png")
```
The reason this method is better than the pandoc approach ![your image](path/to/image)
:
- It automatically changes the command based on the output format (HTML/PDF/Word)
- The same syntax can be used to the size of the plot (
fig.width
), the output width in the report (out.width
), add captions (fig.cap
) etc. - It uses the best graphical devices for the output. This means PDF images remain high resolution.
Tables
knitr::kable()
is the best way to include tables in an R Markdown report as explained fully here. Again, this function is intelligent in automatically selecting the correct formatting for the output selected.
```{r table}
knitr::kable(mtcars[1:5,, 1:5], caption = "A table caption")
```
If you want to make your own simple tables in R Markdown and are using R Studio, you can check out the insert_table
package. It provides a tidy graphical interface for making tables.
Achieving custom styling of the table column width is beyond the scope of knitr
, but the kableExtra
package has been written to help achieve this: https://cran.r-project.org/web/packages/kableExtra/index.html
Style Tips
The R Markdown cheat sheet is still the best place to learn about most the basic syntax you can use.
If you are looking for potential extensions to the formatting, the bookdown
package is also worth exploring. It provides the ability to cross-reference, create special headers and more: https://bookdown.org/yihui/bookdown/markdown-extensions-by-bookdown.html
Several sites provide reasonable cheat sheets or HOWTOs for tables and images. Top on my list are:
Pandoc readme, specifically tables
RStudio's RMarkdown, more details in basics (including tables) and a rewrite of pandoc's markdown.
Pictures are very simple to use but do not offer the ability to adjust the image to fit the page (see Update, below). To adjust the image properties (size, resolution, colors, border, etc), you'll need some form of image editor. I find I can do everything I need with one of ImageMagick, GIMP, or InkScape, all free and open source.
To add a picture, use:
![Caption for the picture.](/path/to/image.png)
I know pandoc supports PNG and JPG, which should meet most of your needs.
You do have control over image size if you are creating it in R (e.g., a plot). This can be done either directly in the command to create the image or, even better, via options if you are using knitr (highly recommended ... check out chunk options, specifically under Plots).
I strongly recommend perusing these tutorials; markdown is very handy and has many features most people don't use on a regular basis but really like once they learn it. (SO is not necessarily the best place to ask questions that are answered very directly in these tutorials.)
Update, 2019-Aug-31
Some time ago, pandoc
incorporated "link_attributes" for images (apparently in 2015, with commit jgm/pandoc#244cd56). "Resizing images" can be done directly. For example:
![unchanged image](foo.jpg)
![much-smaller image](foo.jpg){#id .class width=30 height=20px}
![half-size image](foo.jpg){#id .class width=50% height=50%}
The dimensions can be provided with no units (pixels assumed), or with "px
, cm
, mm
, in
, inch
and %
" (ref: https://pandoc.org/MANUAL.html, search for link_attributes
).
(I'm not certain that CommonMark has implemented this, though there was a lengthy discussion.)
In March I made a deck presentation in slidify, Rmarkdown with impress.js which is a cool 3D framework. My index.Rmd
header looks like
---
title : French TER (regional train) monthly regularity
subtitle : since January 2013
author : brigasnuncamais
job : Business Intelligence / Data Scientist consultant
framework : impressjs # {io2012, html5slides, shower, dzslides, ...}
highlighter : highlight.js # {highlight.js, prettify, highlight}
hitheme : tomorrow #
widgets : [] # {mathjax, quiz, bootstrap}
mode : selfcontained # {standalone, draft}
knit : slidify::knit2slides
subdirs are:
/assets /css /impress-demo.css
/fig /unnamed-chunk-1-1.png (generated by included R code)
/img /SS850452.png (my image used as background)
/js /impress.js
/layouts/custbg.html # content:--- layout: slide --- {{{ slide.html }}}
/libraries /frameworks /impressjs
/io2012
/highlighters /highlight.js
/impress.js
index.Rmd
A slide with image in background code snippet would be in my .Rmd:
<div id="bg">
<img src="assets/img/SS850452.png" alt="">
</div>
Some issues appeared since I last worked on it (photos are no more in background, text it too large on my R plot) but it works fine on my local. Troubles come when I run it on RPubs.