PRESENTATION OVERVIEW

In this presentation...

• Introduce the components of the R toolchain for data science

  • The R Project for Statistical Computing

  • The RStudio IDE

  • \(\LaTeX\)

  • R Tools (for Windows machines)

  • Git/GitHub

• Illustrate the benefits of using this toolchain

• Outline how to install and configure each of the components in the toolchain

  • On personal machines (Windows/Mac/Linux/Unix)

  • On AFIT-networked machines

The R Toolchain for Data Science

First, Let's Define Some Terms

• What is Data Science?

  • Data Science, much like Systems Engineering is a ubiquitous term used to describe any process to convert data into consumable, actionable information.

  • Can be thought of as the intersection of statistics and computer science

  • The rate of progress in statistics has grown SLIGHTLY over the past decade

  • The rate of progress in computer science has grown SO FAST over the past decade that it's impossible to keep up with

  • Data science began when statisticians started capitalizing on the tools developed by computer scientists

A data scientist is someone who knows more about computer science than the average statistician AND more about statistics that the average computer scientist.

• What is a toolchain?

  • Quite simply, a toolchain is a collection of tools that seamlessly link together to improve a workflow

  • Example: The suite of Microsoft Office tools is a toolchain

  • While Microsoft's toolchain is very approachable for the average person - it is FAR less capable that open-source toolchains

  • Microsoft is tightening the integration between MS Office tools, in part by making huge investments in open source tools

  • Building a fully open-source toolchain can be daunting, which is why I'm going to walk you through it!

And unlike proprietary-source tools, you can use this toolchain anywhere

• Each component in this toolchain is

  • Free

  • Open-source

  • Available for all operating systems

  • Available for download on AFIT machines

  • Available for students to use after they leave AFIT

Components of the R Toolchain

The R Project For Statistical Computing

• A general-purpose, interpreted computing language and suite of statistical operators for calculations on vectors, matrices, and arrays

• A large, coherent and integrated collection of tools and graphical facilities for data analysis and reproducible research

• A well-developed programming language which includes conditionals, loops, user-defined recursive functions and input and output facilities

• Has more than 10,000 add-on packages available at the CRAN, GitHub, and BioConductor repositories

• One of the fastest growing technical programming languages in the world

The RStudio IDE

• An integrated development environment (IDE) providing a seamless interface to many open-source computing languages

  • R (*.R)

  • Python (*.py)

  • Julia (*.julia)

  • C++ (*.cpp)

  • JavaScript (*.js)

  • JSON (*.json)

  • XML (*.xml)

  • CSS (*.css)

  • HTML₅ (*.html)

  • TeX (*.tex)

  • YAML (*.yml)

  • Markdown (*.md)

• RStudio also provides a seamless interface to many other open-source applications

  • Pandoc

  • MathJax

  • Git/GitHub

  • SVN

  • Shinyapps.io

  • Apache Spark

\(\LaTeX\) - A complete installation is required to build pdf documents

• MikTeX for Windows 7/8/10

• MacTeX for Mac OS X

• TeX Live 2013+ for Linux

Rtools - A set of compilers for building R packages (Windows only)

• Linux, Unix, and Mac come with a suite of compilers for working with pre-compiled languages

  • C

  • C++

  • FORTRAN

• Windows does not come with these tools so they must be added

  • Ordinarily, these tools would have to be installed one-by-one (and there's a lot of them)

  • Rtools makes life much easier to do this in one step

Git/GitHub

• Git is a free/open-source distributed version control language for ASCII or UTF-8 encoded files

  • Organizes code, text files, images, etc. into repositories 'repos'

  • Easy to collaborate on projects with other users any where in the world without

  • No more emailing files, DVD's, USB drives - It's like dropbox on steroids

  • It can be used on AF NIPR machines (probably not all, but haven't found one yet that couldn't)

  • GitHub has a learning curve, but once you get the hang of it you don't go back

  • Outside of AFIT/AF, GitHub is the standard - some employers prefer GitHub profiles over resumes

"It's become so important...that if GitHub goes down, the software development world practically stops."

• Millions of users - Github.com is currently the 54^th most visited website in the world

• Designed to handle small projects to very large projects with speed and efficiency

  • Writing and sharing software packages

  • Writing books with collaborators from around the world

  • View all changes to the entire volume of German Federal Laws

Why Should You Use This Toolchain (1)?

You can merge ALL of your content into a single file

• Text

• HTML₅, CSS³ and JavaScript code

• Images from local files
• Images from the web
• Videos
• Audio

• HTML Tables

• \(\LaTeX\) tables

\[ \begin{array}{rrrrrrrrrrrr} \hline & mpg & cyl & disp & hp & drat & wt & qsec & vs & am & gear & carb \\ \hline Mazda RX4 & 21.0 & 6 & 160 & 110 & 3.90 & 2.62 & 16.46 & 0 & 1 & 4 & 4 \\ Mazda RX4 Wag & 21.0 & 6 & 160 & 110 & 3.90 & 2.88 & 17.02 & 0 & 1 & 4 & 4 \\ Datsun 710 & 22.8 & 4 & 108 & 93 & 3.85 & 2.32 & 18.61 & 1 & 1 & 4 & 1 \\ Hornet 4 Drive & 21.4 & 6 & 258 & 110 & 3.08 & 3.21 & 19.44 & 1 & 0 & 3 & 1 \\ Hornet Sportabout & 18.7 & 8 & 360 & 175 & 3.15 & 3.44 & 17.02 & 0 & 0 & 3 & 2 \\ Valiant & 18.1 & 6 & 225 & 105 & 2.76 & 3.46 & 20.22 & 1 & 0 & 3 & 1 \\ \hline \end{array} \]

• And interactive tables

• Plots and analysis results with code chunks

Figure 2.6 - Likelihood contributions for different kids of censoring

x <- seq(0,2.4,by = .01)
y <- dweibull(seq(0,2.4,by = .01),shape = 1.7,scale = 1)
plot(   x = x, 
        y = y, 
     type = 'l', 
      lwd = 1.25, 
     xlab = 't', 
     ylab = 'f(t)', 
      las = 1)
polygon(  x = c(seq(0,0.5,.01),0.5),
          y = c(  dweibull(seq(0,0.5,.01),shape = 1.7,scale = 1),0), 
        col = 1)
polygon(  x = c(1,seq(1,1.5,.01),1.5),
          y = c(0,dweibull(seq(1,1.5,.01),shape = 1.7,scale = 1),0), 
        col = 1)
polygon(  x = c(2,seq(2,2.4,.01),2.4),
          y = c(0,dweibull(seq(2,2.4,.01),shape = 1.7,scale = 1),0), 
        col = 1)
text(x = .16,y = .75,'Left Censoring')
text(x = 1.3,y = .65,'Interval Censoring')
text(x = 2.2,y = .15,'Right Censoring')

Why Should You Use This Toolchain (2)?

Compile one input file to many different output formats

• W3C Slidy presetation

• Google ioSlides presentation

• \(\LaTeX\) Beamer presentation

• reveal.js presentation devtools::install_github('rstudio/revealjs')

• shwr.me presentation devtools::install_github('MangoTheCat/rmdshower')

• HTML documents

• Microsoft Word documents

• PDF documents

• Tufte handouts

• Package vignettes

• Journal of Statistical Software article install.packages('rticles')

• Elsevier journal article install.packages('rticles')

• R Journal article install.packages('rticles')

Why Should You Use This Toolchain (3)?

Add Animated plots and Interactive apps

• Static plots and static tables are quickly becoming outdated - like transparencies

• Plots & tables are now expected to provide multiple layers of information

  • Animated plots illustrate complex processes - without having to "visualize" the changes

  • Interactive plots let the viewer change the assumptions and investigate multiple scenarios

  • Interactive tables can be searched and sorted - directly within the document or presentation

• R has two major frameworks to customize how users interact with your presentations

INSTALLING & BUILDING THE R/RSTUDIO TOOLCHAIN

Choose your own adventure

• The links below provide instructions for installing, connecting, configuring, and testing the R/RStudio toolchain for data science

• Note: These steps must be completed for each machine you work on

• It is highly recommended that you build the toolchain on a personal machine or a machine you will have consistent access to (i.e. maybe not a lab machine)

• Follow the link that corresponds to your machine's operating system

  • Personal Windows Machines

  • Personal Mac Machines

  • Personal Linux Machines In Work

  • Personal Unix Machines In Work

  • AFIT-Networked Machines

Still Confused?

But, I'm still having issues!

• Contact me

  • 937.255.3636 ext. 4676

  • jason.freels@afit.edu

My toolchain IS set - Now what?

• If you are new to R, check out these presentations

  • What is R?

  • R Environments and Objects

• If you are familiar with R, this presentation will get you started on how to create documents & presentations and add interactivity

  • Tips and Tricks with rmarkdown

  • Intro to Shiny Apps & HTML Widgets