I put together a guide to how to set up your Linux machine for quantitative work. It is based loosely on the guide put together by Jake Bowers and Jeff Gill available here.

Much (most? all?) of the guide applies to Macs and Windows based machines, although the installation instructions may differ. I think that Ubuntu comes with some developer tools that Mac and Windows machines do not, so if you get an error when trying to run something, that’s where I’d start.¹

The guide is available on Github, and I have copied the readme file here. I encourage you to go to the github repository, as it may be updated and differ from the text here. If you notice anything missing or run into any issues, feel free to use Github’s issues system (or better yet set a pull request) noting the change.

Configuring your Ubuntu-based PC for Quantitative Social Science Work

This file will help you get your Ubuntu-based Linux machine up and running to do quantitative social science data work. It was inspirited by the similarly-titled SocialScienceMacConfig put together by Jake Bowers and Jeff Gill.

Steps should be similar or identical in other Debian-based Linux distros. For other desktop GNU/Linux systems, you’ll probably have access to everything I describe here, but may need to change the commands a bit.

Most (all?) programs included in this list are FOSS. Thus, they are also available on Windows and Mac. Just google around to find out how to get them for your operating system.

Step 0: Install Ubuntu

This is not meant as a manual to help you install Ubuntu. However, you can install Ubuntu on pretty much any computer. I use the 6-month rolling releases instead of the LTS. You can get more information about how to install Ubuntu from Ubuntu’s website. A quick Google search will also pull up helpful instructions. Be sure to test it out from a bootable USB or other disc before installing it. This’ll make sure that it works well on your computer.

Step 1: Install programs from Ubuntu’s repositories

You’ll want Emacs (at least 24) for LaTeX, markdown, R, and git. Yes, it really does all of that. You’ll also want ESS (Emacs Speaks Statistics), which provides a way for Emacs and R to cooperate. Finally, you’ll want git for all your version control needs.

sudo apt-get install emacs24 ess git

If you’re doing Bayesian statistics and want to be able to use JAGS, you’ll also need to get JAGS, which is available and up-to-date on Ubuntu’s repositories: sudo apt-get install jags

Step 1a: Configure git

You’ll need to tell git your name and email. Run these two commands in the terminal, replacing the “John Doe” information with your name and email:

git config --global user.name "John Doe"
git config --global user.email johndoe@example.com

The global option will tell git to remember those values, so you shouldn’t have to do this again.

Step 2: Install R

Installing R on an Ubuntu type Linux distribution is easy. You can simply follow the steps here. Be sure to read about the secure APT part. Note that you can get R from Ubuntu’s repositories, but you really shouldn’t. The version there is pretty dated (3.0 as of this writing, and the latest R version is 3.2.2)

Step 2a: Configure CRAN mirror

This step isn’t totally necessary, but it may save you some time in the future. In your .Rprofile file (usually located at ~/), you can set the CRAN mirror so that R doesn’t ask you. I use RStudio’s mirror, since it should work fairly quickly anywhere in the world. If you want to use a different mirror, replace "https://cran.rstudio.com/" with your favorite CRAN mirror. You can find the full list of mirrors here.

local({
  r <- getOption("repos")
      r["CRAN"] <- "https://cran.rstudio.com/"
          options(repos = r)
})

Step 2b: Install R packages

R can do a lot out of the box, but it’s real strength lies in packages that extend it. Here’s a quick list of some of the packages that I find myself using frequently:

install.packages("lintr", "ggplot2", "dplyr", "zoo",
                 "rstan", "rjags", "MCMCpack", "rmarkdown",
                 "knitr", "reshape2", "servr")

Step 3: Install TeXLive

This is a bit trickier than R. Ubuntu actually has TeXLive in its repositories, but the version there tends to be a bit out of date. If you want, you can install it via sudo apt-get texlive-full. Be warned - it’s a huge file, so it will take a lot of data (and time) to download and install.

I really don’t recommend doing that, though. You can install what’s referred to as “vanilla” TeXLive fairly easily. There’s a stackoverflow answer here that gives step-by-step instructions for how to get vanilla TeXLive set up on your system. In addition to getting the most up-to-date version of LaTeX, the vanilla version will also let you use TeXLive’’s package management system.

If you usually have regular access to the internet, I’d recommend skipping the doc and source files. They add a lot to the installation and are totally unnecessary if you have internet connectivity.

If you’re on a Mac, you can install MaCTeX and Windows users probably want MiKTeX.

Step 4: Configure Emacs

Emacs is infinitely configure-able. You could spend months (years?) getting it just right. Feel free to do so. It’s probably not worth it, though. My advice is just to commit to using emacs, and develop your specific configuration as you go. There are definitely a few things that you’ll want to set up beforehand, though.

I keep my .emacs file in a github repository, so feel free to check it out. If you drop the init.el file into ~/.emacs.d/, it should work just fine on your setup. It adds the ability to connect to MELPA to your emacs setup, so you’ll need an internet connection the first time you use it.

I’m not sure I’d recommend just using my .emacs file, though (or anyone else’s). Setting up emacs yourself teaches you a bit about how the program thinks. So I’ll list some packages here that I think are particularly invaluable:

• magit - for all your git needs, inside emacs
• auctex - for writing LaTeX
• polymode - for .Rnw (knitr + LaTeX) and .Rmd (rmarkdown) files
• ebib - for .bib file management
• markdown-mode - for working with markdown files
• reftex - this comes bundled with emacs in recent versions (so no need to install it separately), but makes working with .bib files a breeze

Those are absolute must-haves for doing the sort of work we do. If you work with languages other than R, LaTeX, and markdown, you’ll want to make sure you have the proper “major mode” for that language. ESS (which you already installed above) makes emacs work with R, so no need to do anything there. There are some other packages that just make emacs work a little better, and I’ll list some here:

• better-defaults - literally better defaults for emacs. Pretty much everyone should use this
• smex - easier way to navigate commands (like M-x)
• There are several “ido” packages - go look them up. They’ll make your life easier.
• smartparens - easier parentheses management
• company - for autocompletion (there’s also auto-complete, but I prefer company)
• flyspell - for spell checking on the fly
• smooth-scrolling - for doing away with emacs’ nonsense default scrolling
• flycheck - checks for good style and syntax

There are tons more packages available, but those should get you started.

Step 5: Profit!

You’re now set up to use a combination of git, R, markdown, and LaTeX in a sane working environment. You didn’t have to pay a penny, either! These programs (emacs and R in particular) have a bit of a learning curve, but don’t be dissuaded. There are tons of resources online that can help you out. I also write about this kind of stuff on my blog.