README.md

In pheymanss/chronicle: Grammar for Creating R Markdown Reports

{chronicle}

2021-06-24

An R package for easy R Markdown reporting

install.packages('chronicle')

This R package aims to be an opinionated assistant, to whom you can delegate the task of creating visually appealing and consistent R Markdown reports.

A quick demo

The way you build the reports is by specifying the structure of your report through the add_* family of functions, layering one below the previous one.

library(chronicle)

demo_report <-
  add_text(text_title = "This is the output of a chronicle call",
           text = "Each element has been added through an `add_*` function.",
           title_level = 1) %>%
  add_table(table = head(iris),
            table_title = "A glimpse of the iris dataset",
            html_table_type = "kable",
            title_level = 1) %>%
  add_raincloud(dt = iris,
                value = "Sepal.Length",
                groups = "Species") %>%
  add_scatterplot(dt = iris,
                  x = "Petal.Width",
                  y = "Petal.Length",
                  groups = "Species")

render_report(report = demo_report,
              title = "A quick chronicle demo",
              filename = "quick_demo",
              keep_rmd = TRUE)

You can see the output of this call, and a full showcase of all the elements supported by chronicle.

What happens behind these calls is that chronicle literally wirtes the content of an R Markdown for you! You can see the content of the report by directly printing it.

demo_report

 # This is the output of a chronicle call

 Each element has been added through an `add_*` function.

 # A glimpse of the iris dataset
 ```{r, echo=FALSE, message=FALSE, warning=FALSE}
 knitr::kable(head(iris))
 ```

 ## Distribution of Sepal.Length by Species
 ```{r, echo=FALSE, message=FALSE, warning=FALSE, fig.width=params$figure_width, fig.height=params$figure_height}
 chronicle::make_raincloud(dt = iris,
                           value = 'Sepal.Length',
                           groups = 'Species',
                           adjust = 0.5,
                           include_boxplot = TRUE,
                           include_mean = FALSE,
                           include_median = TRUE,
                           force_all_jitter_obs = FALSE,
                           ggtheme = 'minimal',
                           plot_palette = params$plot_palette,
                           plot_palette_generator = params$plot_palette_generator,
                           static = params$set_static)
 ```

 ## Petal.Width vs Petal.Length by Species
 ```{r, echo=FALSE, message=FALSE, warning=FALSE, fig.width=params$figure_width, fig.height=params$figure_height}
 chronicle::make_scatterplot(dt = iris,
                             x = 'Petal.Width',
                             y = 'Petal.Length',
                             groups = 'Species',
                             plot_palette = params$plot_palette,
                             plot_palette_generator = params$plot_palette_generator,
                             static = params$set_static)
 ```

The make_* family of functions

Every plot added with an add_* function will be built through its correpsonding make_* function. These functions take care of the heavy lifting, avoiding the cumbersome (albeit powerful) sintax of ggplot, plotly and other html widgets. The parameters of the make_functions are simple and intuitive specifications on how to make each plot, and they can be called independently and used in any instance where a ggplot or an html widget would fit.

make_barplot(dt = ggplot2::mpg,
             value = 'cty',
             bars = 'manufacturer',
             break_bars_by = 'cyl',
             horizontal = TRUE,
             sort_by_value = TRUE,
             static = TRUE)

make_raincloud(dt = iris,
             value = 'Sepal.Length',
             groups = 'Species')

Rendering chronicle reports

Once the structure of the report has been defined, the rendering process is done by render_report(). This uses rmarkdown::render() as a backend for rendering the report, which gives chronicle the capability to render the reports with full visibility to all objects in the global environment. This gives chronicle two of big strengths:

1. You don't need to include nor run all your data processing code again for a new report output. This means you can build several report recipes for different audiences out of the same data processing, rendering all of them from the same global environment.

2. It can render several output formats in a single call. For instance, it is possible to render the same content as ioslides for a presentation, as tufte_html for handouts and as rmdformats for a site upload.

Take our quick demo as an example, to render this as the three outputs mentioned previously, you only need to add that vector to the output_format parameter of render_report()

render_report(report = demo_report,
              output_format = c("ioslides", "tufte_html", "rmdformats"),
              filename = "quick_demo",
              title = "A quick chronicle demo",
              author = "You did this!",
              include_date = FALSE,
              fig_width = 8,
              fig_height = 6,
              plot_palette = c("#FC8D59", "#FFFFBF", "#99D594"),
              plot_palette_generator = "plasma",
              rmdformats_theme = "readthedown",
              keep_rmd = TRUE)

The render_report function allows the user to specify several global parameters of the report, like fig_width, fig_height, plot_palette, and plot_palette_generator. This last one specifies which palette from the viridis package will be used to complete insufficiently long palettes (or when the user did not specify a palette at all.)

make_violin(dt = iris,
             value = 'Sepal.Length',
             groups = 'Species',
             plot_palette_generator = 'mako')

make_density(dt = iris,
             value = 'Sepal.Length',
             groups = 'Species',
             faceted = FALSE,
             plot_palette_generator = 'turbo')

The report_columns() function

chronicle also includes a function called report_columns(), that will create an entire chronicle report for a single dataset. It includes a comprehensive summary of the data through the skimr::skim() function, along with one plot for each column present in the data: bar plots for categorical variables and rain cloud plots for numerical variables. This gives you an immediate view of a dataset with a single line of code!

report_columns(dt = palmerpenguins::penguins,
               by_column = 'species')

And you can see the output of this call

Supported formats

As of version 0.2.5, chronicle can output both static and dynamic outputs. Dynamic outputs refer to R Markdown formats that support html widgets, hence the elements added will be dynamic plots (plotly, dygraph, DT). For static outputs, these will roll back to ggplot and static table prints. To avoid large file sizes, it will also roll back to static plots if the table has over 10,000 rows.

Dynamic outputs (html)

• bookdown
• github_document (you are reading one right now!)
• html_document
• html_notebook
• ioslides
• prettydoc
• rmdformats (this is currently the default)
• slidy_presentation
• tufte_html

Static outputs

• pagedown
• pdf
• powerpoint_presentation
• rolldown
• tufte_handout
• word_document

Additionally, {flexdashboard} and {xaringan} technically compile, but the layout is stiff in flexdashboard and altogether incorrect in xaringan, so you might still use it if you don't mind manually correcting the separators. Also, {rticles} support can technically be added, but that would involve a plethora of additional parameters for the header, and frankly, writing a journal article is not the intended use of the package

Supported report elements

I highly encourage you to review the enitre showcase, words are not as adequate to describe each element. But for a quick glance, as of version 0.2.5 chronicle supports:

• Bar plots
• Box plots
• Code (optionally evaluated)
• Density plots
• Dygraphs
• Histograms
• Line plots
• Rain cloud plots
• Scatter plots
• Tables
• Texts
• Titles
• Violin plots

pheymanss/chronicle documentation built on Jan. 19, 2024, 6 a.m.