Dear all,

This week we’ll extend the knowledge we have about rmarkdown to generate presenations in html and pdf format. We’ve seen in week 4 that RStudio’s markdown implementation is a very quick and convenient language to quickly mark up text documents and, if necessary, to include images, mathematical equations (\(\LaTeX\)) and to display and/or execute code (R, Python, etc).

Let’s take rmarkdown a step further.

Best,

Gerko

1 rmarkdown for presentations

Generating presentations with markdown is very much like marking up regular documents in markdown. However, there are some minor differences:

First, presentations usually follow some form of dimensionality that adheres to display conventions: screen dimensions are historically obeying to 4:3 or 16:9 aspect ratios, although recent trends do allow for wider 18:9 or 21:9 aspect ratios. The benefit of marking up regular html documents with rmarkdown is the natural scaling of content to the device. With pdf’s the resulting document is paginated and follows a strict canvas size (e.g. A4, US letter, etc) suitable for printing.

Second, output options need to be explicitly formulated. There are multiple implementations for generating html presentations and, when the output is set to pdf, the beamer specifications have to be declared.

Third, the pound sign # is still used for (sub)sections, but the leveling is very important: a single # indicates a new chapter in your presentation, a double ## indicates a new slide within that chapter.

1.1 Different forms of output

RStudio supports the following slide output styles:

  • html ioslides: My preferred mode of creating presentations. See this page for a detailed how-to.
  • html slidy: Another flexible mode of creating presentations. See this page for a detailed how-to.
  • pdf beamer. See this page for a detailed how-to.

I suggest you take a look at these output styles and pick a style. You can always switch later on and have your rmarkdown document generate a different style.

1.2 References

See this page on a detailed walkthrough about how rmarkdown handles bibliographies and citations.

1.3 Interactive graphs

With package plotly, any graphics generated by ggplot2 can become interactive in html output files. It is adviced to install the developer version of ggplot2 from GitHub. You can do so by running

devtools::install_github('hadley/ggplot2')

An example:

require(plotly)
set.seed(123)

df <- data.frame(x <- rchisq(1000, 5, 10),
                 group <- sample(LETTERS[1:5], size = 1000, replace = T))

p <- ggplot(df, aes(x, fill = group)) + 
  geom_density(alpha = 0.5, position = "stack") + 
  ggtitle("stacked density chart")

ggplotly(p)

1.4 Interactive tables

With package DT, any table in R can become interactive in html output files. An example that outputs the first 5 rows by default:

require(DT)
datatable(mtcars, options = list(pageLength = 5))

1.5 FlowCharts and Diagrams

DiagrammeR it is straightforward to generate flow charts from within any rmarkdown document. An example flowchart:

library(DiagrammeR)
grViz("
  digraph {
    layout = twopi
    node [shape = circle]
    A -> {B C D} 
    B -> {C D}
    D -> {A B}
  }")

Or, more complicated:

grViz("
  digraph dot {

    graph [compound = true, nodesep = .5, ranksep = .25,
           color = crimson, label='Polynomial Regression Model'
           /*, rankdir='LR', style=filled, fillcolor = blue*/
          ]

    subgraph cluster1 {
      node [shape = diamond, 
            color = black]
      xi

      node [shape = circle,
            style = filled,
            fillcolor = grey] 
      ti
      
      edge [color = black] 
      xi -> ti

      label='N'
    }  
  
    node [shape = circle]
    w

    node [shape = diamond,
          color = black,
          label = '&alpha;']
    alpha

    node [shape = diamond,
          color = black,
          label = '&sigma;&#x00B2;']
    sigma

    edge [color = black]
    alpha -> w
    w -> ti
    sigma -> ti 

  }
  ",
engine = "dot")

An example diagram sequence with DiagrammeR:

DiagrammeR("
sequenceDiagram;
   customer->>ticket seller: ask ticket;
   ticket seller->>database: seats;
   alt tickets available
     database->>ticket seller: ok;
     ticket seller->>customer: confirm;
     customer->>ticket seller: ok;
     ticket seller->>database: book a seat;
     ticket seller->>printer: print ticket;
   else sold out
     database->>ticket seller: none left;
     ticket seller->>customer:  sorry;
   end
")

1.6 htmlwidgets for R

Have a look at the htmlwidgets website to lear all about referencing, tweaking, linking, using and misusing a variety of html widgets in rmarkdown.

2 Exercise

Create a presentation about any topic you like that consists of at least 7 slides and includes all of the following:

  • a logo
  • a centered still figure
  • an interactive table
  • a moving figure, interactive figure or movie
  • a 2-column slide
  • an aligned multi-row equation
  • a citation and reference list
  • r-code, displayed but not executed
  • cached r-code (may overlap with the next requirement)
  • r-code, executed, but not displayed (e.g. a figure generation)

Of course this presentation may serve as the basis for presentations that you may have to create for other courses. Submit the presentation to this year’s GitHub repo.