APHREA: Workshop Shiny

Transcript

1. 1 SHINY Jeff Goldsmith, PhD Department of Biostatistics
2. 2 • Framework for building interactive plots and web applications in R • Shiny allows you to create a graphical user interface (GUI) – Users can interact with your code without knowing R! – Communicate visualizations, models, algorithms to collaborators • Uses HTML, CSS, and JavaScript framework – You don’t need to know these to use Shiny – The syntax can be tricky at first, though – Knowing more can help you get fancy • Brought to you by RStudio in 2012 What is Shiny?
3. 3 • Package for creating web-apps • Don’t need to learn how to code apps directly; you write R code and shiny creates then app – Analogous to creating HTML files by writing R Markdown and knitting • Adds interactivity – your app can take user input and update outputs accordingly • For a quick example, run shiny::runExample("01_hello") in your R console What is Shiny?
4. 3 • Package for creating web-apps • Don’t need to learn how to code apps directly; you write R code and shiny creates then app – Analogous to creating HTML files by writing R Markdown and knitting • Adds interactivity – your app can take user input and update outputs accordingly • For a quick example, run shiny::runExample("01_hello") in your R console What is Shiny?
5. 4 • Shiny documents / applications have two components: – A user interface to obtain inputs – Code that reacts to inputs and produces outputs • R code executes in the background • Because you need R to use Shiny, sharing Shiny-based products requires some thought – Not as “easy” as sending / hosting HTML files produced only by R Markdown How does Shiny work?
6. 5 • Widgets are text elements that users can interact with – Examples include scroll bars, buttons, text, ect – Take in user input Getting inputs
7. 6 • These are functions that react to user input from widgets – renderPrint() -- prints output of a function – renderText() -- outputs text – renderTable() -- for making tables – renderPlot() -- outputs plot made using ggplot2 (and base R, …) – renderPlotly() -- outputs plot made with plotly library Producing outputs
8. 7 • R-Markdown-based Shiny document • Relatively easy to use (given an understanding of dashboards / markdown) • Adds dynamic elements to a flexdashboard – Input / output elements are added directly to the R Markdown file Flexdashboard + Shiny
9. 8 • Standalone Shiny app • Not built within an R Markdown document – Separate .R files control UI and “server” computations for input / output – Alternatively, UI and server objects included in a single app file • Potentially more flexible than piggybacking on R Markdown / flexdashboard Shiny applications
10. 9 Shiny applications • ui – Controls layout and appearance – Where you add widgets – ui.R • server – Instructions your computer needs to build the app – R code for plots, etc – server.R
11. 10 • Not always easy – Shiny requires R in the background to update outputs based on inputs – HTML files can’t do this • Providing files – Send “raw” files (.rmd, .R, data, etc), maybe as an R project – Recipient knits the file / runs the app through Rstudio • Hosting online – Needs a server that runs R in the background, and github doesn’t – shinyapps.io is pretty common way to permanently host document / app Sharing shiny products