shinyData - GUI for data analysis and reporting

Written on 2015-03-19

Some people find very hard to start using R because it has no GUI.
There exists some GUIs which offers some of the functionality of R.
In this post I would like to focus on one such GUI, a very new shiny application called shinyData.
I hope the app will make it easier for some to get into R environment.
Also it can reduce development time of analysis and reports for existing R users.


The shiny app is well described in this post already. In my post I will make a live test of it on my sales data.
App is hosted on slow so I would recommend to run it locally. It requires/recommends particular version of dependencies, I will use new library:

lib <- paste(getwd(),"shinyData_lib",sep="/")
  install_github("ebailey78/shinyBS", ref = "shinyBS3")

shinyData live

Once you've launched app you can see a projects page. You can choose some examples, I will skip to data tab as I prefer to load own data from csv file. Yet very useful feature is ability to save project and load it later. I highly recommend to use that feature as it costs nothing and may save your work.


I have csv of randomly populated sales data dimensioned by currency, geography and time. Once I've loaded data I see its preview. App will automatically recognize measures in my dataset.



The most interesting part is to produce some information from the data.
First I need to choose the data, then I can start to do mapping between columns from dataset to the elements on plot, particularly the ggplot2.
You can make multiple sheets of visualization, I will create two.

I've put time on X axis and value on Y axis, values are grouped into division name using colors and organized into the panels by region and currency type


Results can looks like


One more plot with year on X axis grouped into division names, value on Y axis



I have my two sheets of visualizations so I can try to build some report based on them.
Presentation tab combines a plaintext (markdown) editor and gives ability to nest elements from R, in this case also the products of visualization.
This is simply what rmarkdown package do, yet you don't need to know rmarkdown. Write the text as in the notepad and add your visualization sheets by click on Insert Sheet element on the left sidebar.

Below is the content of the report that I'm going to produce.
Important to note: rmarkdown gives ability to render your documents into various format: html, pdf, word, markdown, etc.



shinyData is a well built shinyApp which can make the life easier for some R user, and for new ones limit the learning curve to minimum still giving possibility to produce some nice reports.
Personally I prefer to work on fully reproducible scripts, so for me shinyData is more a presentation of capabilities which R, shiny and open source community can bring.