Literate programming

• Combine text, code, figures, tables, etc.
• Write papers without having to remember to swap figures
• Automatically generate reports

Getting started

• Open RStudio
• File -> New File -> R Markdown
  • Enter a title and author(s).
  • Choose ‘Default Output Format’ as HTML.
• Generates a basic stub of a .Rmd document
  • Customize “front matter” at the top of the document between the ---.
• Press Knit to create an HTML from the document
  • 1st time may ask you to install some packages
  • install.packages("rmarkdown") in console for permanence
• Runs the code in the code chunks and prints their output along with the markdown formatted text
• Can also create PDF & Word versions of our files
  • PDF Requires pandoc and TeX installation
  • Use the Knit dropdown or change output: pdf_document
• R Notebook
  • output: html_notebook
  • “Interactive R Markdown”
    • execute individual code chunks
    • output in editor pane

Markdown

• Basic approach to formatting text
• Let’s you do
  • *italics*
  • **bold**
  • [links](http://google.com)
  • Lists
    • *
    • 1.
  • # Header
• Easy to read for humans
• Easy to convert into other things for computers
• Common on lots of websites
• Used to create all of the exercises and lectures in this course
• Github will automatically render it
  • https://github.com/ethanwhite/CV/blob/master/CV.md

Explore patterns in population dynamics at Portal.

## Required Libraries

R chunks

• Set R code inside a set of ``` with the {r} designation

```{r}
```

• Code that you write inside chunks gets executed during the “knit” process and the results are shown below.

```{r}
library(dplyr)
```

• Chunks have lots of useful options
  • Options are described at: http://yihui.name/knitr/options/
  • Options will be listed in RStudio if you press tab inside the {r} brackets at the top of the chunk

```{r, message=FALSE}
library(dplyr)
```

```{r, message=FALSE}
library(dplyr)
library(ggplot2)
```

```{r, message=FALSE, warning=FALSE}
library(dplyr)
library(ggplot2)
```

• You can run code inside your text, too:
  • `r cos(pi)` turns into -1 when you press the Knit button.
  • We will see and example of this later.

Analysis Report Example

• Here’s a text segment linked to a code chunk that begins an analysis report.

## Data

Data is from the [Portal project teaching database](http://figshare.com/articles/Portal_Project_Teaching_Database/1314459) 
published on Figshare. We need the surveys table for our analysis:

```{r, cache=TRUE}
download.file("https://ndownloader.figshare.com/files/2292172", 
              "surveys.csv")
data <- read.csv("surveys.csv")
```

• Setting cache=TRUE lets you reuse the results of this code chunk in subsequent “knits” instead of re-calculating or re-downloading it each time.

## Analyze population time-series

Get the time-series of counts for all species.
          
```{r}
time_series <-
  data %>%
  group_by(species_id, year) %>%
  summarize(count = n()) %>%
  na.omit()

head(time_series)
```

• echo=FALSE let’s you show the results of the code chunk without showing the code.

## Plot the time-series.

```{r, message=FALSE, echo=FALSE, cache=TRUE}
ggplot(time_series, aes(x = year, y = count)) +
  geom_point() +
  geom_line() +
  facet_wrap(~species_id) +
  scale_x_continuous(breaks = pretty_breaks(n=2))
```

## A simple model

```{r, echo=FALSE}
model <- data %>% group_by(year) %>% 
  summarize(count = n()) %>% 
  lm(count ~ year, data = .)

results <- anova(model)
```

• Here’s the example of an inline code chunk.

We found a marginally significant linear relationship between the
total count and year (p = `r round(results[["Pr(>F)"]][1], 3)`; see
Table 1 for more details)

```{r, echo=FALSE}
knitr::kable(results, caption = "Table 1")
```

• knitr::kable() is a handy way to make nice-looking tables from data frames.

R Presentations

• File -> New File -> R Presentation
  • Save to location.
  • Rpres generates MD and directory of figures.
• Template Rpres file includes title page and sample code blocks.

Untitled
========================================================
author: 
date: 
autosize: true

First Slide
========================================================

For more details on authoring R presentations please visit <https://support.rstudio.com/hc/en-us/articles/200486468>.

- Bullet 1
- Bullet 2
- Bullet 3

Slide With Code
========================================================

```{r}
summary(cars)
```

• Preview in RStudio panel
• Convert to HTML
  • More -> View in Browser or Save As Web Page...