Creating dashboards using "flexdashboard" R package

March 04, 2019

Creating a dashboard is an attractive way to visualize different groups of related data. To setup a dashboard we can use the R package flexdashboard.

First, setup the orientation of the dashboard in YML header. The default orientation is columns, which shows individual charts stacked vertically within each column. To setup the orientation row-wise specify orientation: rows option in YML header.
Similarly, we can display several components in different windows using a tabset. To add a tabset layout as a row or column we have to add the {.tabset} attribute to the section heading. The widths and heights of rows and columns can be set by using data-width and data-height attributes when we define a row or a column.
In the dashboard, we can include a wide variety of components.

One such component is htmlwidgets, for which there are over 30 packages on CRAN. Some of the available htmlwidgets are Leaflet, dygraphs, Plotly, rbokeh, Highcharter, and visNetwork. For larger data sets, use standard R graphics instead of htmlwidgets.
We can use any chart created with standard R graphics (base, lattice, grid, etc.) with flexdashboard.
Tabular data can be included within flexdashboards as a simple tabular display which is suitable for smaller number of observations,or as a DataTable. For simple tabular display, we can use the knitr::kable function to display the dataset. For DataTable, we can use DT package to display matrices or data frames as interactive HTML tables.
We can also use the valueBox function to display single values along with a title and optional icon, and gauge function to display values on a meter within a specified range.
Including a navigation bar is also possible to the flexdashboard. By default it shows document’s title, author and date. In addition to that, we can add custom links and social links as well.
If we need to include additional text in the dashboard, we can add the content at the top of the page before dashboard sections are introduced, or we can add a dashboard sections which don’t contains charts, but only text, images equations etc.

If We expect to divide the dashboard into multiple pages, define a page by adding a level 1 markdown header (==================). For each page we can have different page orientation by adding the data-orientation attribute to the header of the page. e.g. page 1 {data-orientation=rows}
To access a page via a navigation bar menu, we can use the data-navmenu attribute as below: Page 1 {data-navmenu="Menu 1"}
Also, we can link different pages or hide a page using options [Page Two](#page-2) and Page 4 {.hidden}, respectively.
Further, we can add an icon to the navigation menu for a page using the data-icon attribute.
e.g. Page 2 {data-icon="fa-list"}
The icons can be selected from three different icon sets, Font Awesome, Ionicons, and Bootstrap Glyphicons.
Using flexdashboard package we can also create a storyboard layout. To create a storyboard, add the storyboard: true option in the YML header. Then include a set of level 3 (###) dashboard components as a frame which is a navigation caption.
e.g. ### Frame 1
Further, we can layout one or more pages of a dashboard as storyboards and other pages as usual row and column layout of a dashboard. In this case, we have to exclude the storyboard: true option from YML header, and add the {.storyboard} attribute to pages that we expect to lay out as storyboards.

Some flexdashboards examples are shown here.

More details

R
R Markdown