3  Introduction to R

This article was written by Wan Hsien Heah and originally published on 21 Sep 2020

There are several different data science tools available today. R is one such tool that is used quite widely, free and easy to get started with. R started out as an experiment at the University of Auckland to see if an existing language called S could be improved such that code would be cleaner and easier to run. It has come a long way since its humble beginnings in the mid-90s. Today, R is used for a variety of tasks including complex capital model calculations and running machine learning algorithms in pricing and reserving.

3.1 Why use R?

There are a number of benefits to using R. These are:

  • Free: There is no license fee associated with using R (although beware that there may be license restrictions with using R packages).

  • Widely used: R has a large user base, with data scientists, data miners, statisticians and actuaries using the tool around the world. That means that it only takes a quick search on the internet to get help with any problem that you might be experiencing with coding in R.

  • An R-ready workforce: R is now a part of the UK Institute and Faculty of Actuaries (IFoA) exam syllabus. Increasingly, more and more university students are also learning R as part of the undergraduate course in statistics and actuarial science.

  • R is fast: R is usually faster at computational problems compared to traditional tools like Excel. For example, R can generate a million random standard Normal samples in less than a second. This is because of the way R is structured and the way it works. R can even be faster than server-based languages like SQL.

  • Packages: There are many freely available reusable R code libraries available online called packages. These packages have been built by other R users and perform a variety of calculations or processes which means that you do not have to start from scratch all the time. The crowd-sourced sharing of packages helps accelerate development time in R. We will cover some of the more useful packages for actuaries in our future blog posts.

3.2 How do I get started?

R is available to download from the Comprehensive R Archive Network (CRAN) at https://cran.r-project.org/. CRAN is a network of servers around the world that contains copies of the R application and the various packages which is continuously being updated and maintained. To get started, you will first need to download R from the CRAN homepage.

CRAN homepage

Select the appropriate download for your operating system and follow the usual instructions to compile. You will usually want to download the most recent version, but occasionally you may need to download older versions if, for example, a package you need to use has not yet been updated to run on the latest version of R.

The instructions for each platform will be on the website as you click the version of R that you need. For this blog post, we will focus on the Windows version of R as it is the more common platform used by actuaries.

For first time users, you will need to install base R. It is important to note that base R is a console application i.e. the main interaction is through text. Most users will want to use an Integrated Development Environment (IDE) or a user interface. So, after downloading R, you should consider installing an IDE such as RStudio. You should install R first, then RStudio.

RStudio provides a graphical user interface (GUI) that provides a more visual interaction when using R. You can download a free desktop version from the RStudio website: https://rstudio.com/. There are also paid options for support and for running and maintaining a server.

Once you have both R and RStudio installed, you are ready to begin coding.

3.3 Hello World!

Now that you have the minimum software required installed, you can start coding. Getting started can be quite daunting as there are various schools of thought on how to use R and which packages. Base R is what comes built into R. There are also a growing series of data manipulation packages commonly known as tidyverse which consists of packages developed and maintained by RStudio. Beyond that there are other data manipulation packages such as data.table. Each have their advantages and disadvantages depending on the precise operation that you would like to perform

There are several free resources available for learning R which you can find on the internet. We list a few of them here:

  1. R for Data Science
  2. R-bloggers - an aggregator of online articles about R
  3. Rweekly.org - a weekly curated list of updates from the R community

For more advanced users, useful reference guides are:

  1. Advanced R, first edition more advanced techniques using mainly baseR
  2. Advanced R, second edition the updated book, which incorporates some of the tidyverse ideas.

Additionally, you could also sign up to online courses from Coursera, Stanford, CodeAcademy, etc. that teach R with the added benefit of certification at the end.

3.4 Conclusion

There are, of course, other free statistical tools available out there. We have chosen to focus on R as it is one of the more prevalent programming languages being used in the actuarial space. R brings with it many benefits which you may find useful in the work that you do. Have a go and try it out!