- Use multiple languages including R, Python, and SQL.
- Output formats include HTML, PDF, MS Word, Beamer, HTML5 slides, Tufte-style handouts, books, dashboards, shiny applications, scientific articles, websites, and more.
Rmarkdown Refresher
The Precursor to Notebooks
Problems Rmarkdown Solves
- Combines code and output
- Facilitated by RStudio (but no IDE required)
- Version control!!!
- http://danielphadley.com/Jupyter-to-Rmarkdown/
Branding a Word Doc
Example
Example
Example
Useful Example: Page Breaks
- Set up a docx style reference
- Modify style ‘Heading 5’ :
- Set the font color to ‘white’ (rather than ‘Automatic’).
- Select the smallest font size (8 rather than 11).
- Select ‘Page break before’ in the ‘Line and Page Breaks’ tab.
https://datascienceplus.com/r-markdown-how-to-insert-page-breaks-in-a-ms-word-document/
Example
##### Pagebreak # Overview The raw data comes from ...
Branding Images
Branding Images with Magick
Branding Images with Magick
Branding Images with Magick
# Our function in Pseudocode
si_ggsave <- function(){
ggsave(lastplot)
if(add_logo){
logo <- image_read("~/ourlogo.png") %>%
#make the width of the logo match the width of the plot
image_crop(paste0(pwidth, "x0+0+0"))
# The final version is stacked on top of the sorenson logo
final_plot <- image_append(c(lastplot, logo), stack = TRUE)
# And then we overwrite the standard ggsave call
image_write(final_plot, filename)
}
}
Packaging Your Brand
Customizations Add Up
Setup chunk: over 100 lines
- Load libraries
- Set knitr/rmarkdown options
- Set directories
- Define SI color names
- Customize ggplot themes and geoms
- Define custom functions
Goals
Part 1
- Reduce setup chunk
- Corral custom functions
Part 2
- Effortless creation of new RMarkdown documents
- Single-source templates that upgrade easily
- Future proof / time-locked
Reduce Setup Chunk
Move all settings into broad "apply_settings()" type functions
si_ggplot_theme_apply <- function() {
ggplot2::update_geom_defaults("bar", list(fill = si_design$pacific))
ggplot2::update_geom_defaults("smooth", list(
colour = si_design$pacific,
fill = si_design$arctic, alpha = I(2/10)))
ggplot2::update_geom_defaults("point", list(
colour = si_design$pacific,
fill = si_design$pacific))
ggplot2::update_geom_defaults("col", list(fill = si_design$pacific))
ggplot2::theme_set(ggplot2::theme_minimal())
}
Easy Creation of Rmarkdown Template
- inst/ dir in package can hold anything
- Contents are copied to local R/library/package dir on package install/update
- Current version copied to working dir on each create
Using a Package for Template Storage
- inst/ dir in package can hold anything
- Contents are copied to local R/library/package dir on package install/update
- Current version copied to working dir on each create
Draft Function
# Our draft function in Pseudocode
si_draft_full_report <- function(file) {
if(empty(dirname(file))) { #Create dir if not provided
file <- file.path(default_dir, file)
}
if(tools::file_ext(file) != "Rmd") { #Fix the capitalization of Rmd
message("Daniel you messed up the capitalization again. Setting extension to .Rmd")
file <- paste0(tools::file_path_sans_ext(file), ".Rmd")
}
rmarkdown::draft(file = file,
template = "si_full_report",
package = "sorensonimpact",
edit = F)
file.edit(file) #Open immediately in Rstudio editor
}
Draft Function in Action
Finishing Touches
One final piece…
Finishing Touches
One final piece…
Automating Your Work
An Old Method: Knitting to HTML
<!DOCTYPE html>
<script>
$(function() {
$('#container').highcharts({
xAxis: {
categories: [<!--rinline I(shQuote(dates)) -->]
},
series: [{
name: 'Tokyo',
data: [<!--rinline I(Tokyo) -->]
}, {
name: 'London',
data: [<!--rinline I(London) -->]
}]
});
});
</script>
Automating is Easy Now
for (county in unique(utah$County)) {
try(render('./county_report.Rmd',
output_file = paste(county, "County Report.docx"),
output_dir = "./County_Reports/"))
}
A New Method
Create a script that
- loads data (e.g, from Google Sheets)
- renders a report from an .Rmd
- saves the output (e.g., an HTML dashboard)
- emails the result
Automate the Script
- Windows task manager OR
- Cron (e.g., on a Raspberry Pi)
https://www.r-bloggers.com/scheduling-r-markdown-reports-via-email/