Knowledge, Information and Data (KIND) learning site

This training course is an introduction to programming in R, with particular emphasis on using tidyverse functions effectively. It is designed with data analysts working in health, care, and housing in mind, and has a slight bias towards Rmarkdown as a likely endpoint of the programming activities that it teaches.

The training expects you to have some basic familiarity with R and Rmarkdown, including the tidyverse. While the expectations are fairly low - you absolutely don’t need to be an R expert to successfully complete this training - you should have some previous experience using R. As the course has a tidyverse flavour, trainees should have some familiarity with using core tidyverse functions (dplyr especially).

Some self-assessment questions might be useful to calibrate these expectations. Are you able to answer these questions fairly rapidly and with confidence?

• what’s a tibble?
• what function would you use to load a .csv file?
• how would you make a new column in a tibble from some other column(s)?
• how would you draw a simple graph using ggplot()?

If you were able to answer these fluently, then great, this training is meant for you. If they were a bit beyond your previous experience, then you might like to enroll on the KIND Learning Network training Dynamic Reports in R/Rmarkdown as a first step in your R journey.

This is a modular training course split into five sessions.

1. Getting the best out of dplyr

Going beyond the basics of dplyr (filter(), select(), mutate()). Using the NHSRdatasets as a model system for health and social care analysis, this session introduces dplyr tools for investigating data, gives an introduction to tidyselect, and provides several means of summarising and joining data.

2. Functions

An introduction to writing functions in R. Includes advice about translating existing code into functions, and brief primers on indirection for tidyverse functions, conditional execution, and dot-dot-dot (…).

3. Iteration

An introduction to the often-maligned for-loop in R. Also includes training on vectorised and non-vectorised functions, and gives some examples of using for-loops to create Rmarkdown.

4. The purrr package

An introduction and walk-through of the purrr package, showing how to apply functions to several different kinds of object in R.

5. tidy evaluation

Tidy evaluation is a tidyverse-specific approach to programming. This session introduces two different elements of tidy evaluation: tidyselect, which are tidyverse-specific tools for working with columns, and data masking, which are tools for using data as variables, and vice versa.

You can take any combination of the sessions in any order. Of necessity, there is some overlap between sessions, but in general they are independent and can be mixed-and-matched to suit your interest.

How does it work?

The training is delivered as a series of live Teams session using Rstudio Cloud.

If you’ve never used Rstudio Cloud before, it’s a SaaS version of Rstudio in the browser. Note that because it’s a web service, it requires you to upload your data to their servers, which might makes it unsuitable for production work in health and care owing to information governance concerns. That said, it’s an excellent venue for training, because it solves many of the difficulties regarding R versions, permissions, etc that are a feature of using R on the desktop. It’s also very easy to transfer projects from RStudio Cloud to an installed version of R, so don’t worry that what you learn here will be tied to the cloud forever.

What you’ll need

As RStudio Cloud is a web service, you don’t need a particularly up-to-date computer to completed this training. As long as you have a reliable internet connection, and are capable of making a video call with Microsoft Teams (for the face-to-face part of the training), then you should be fine. The demonstration has been tested on Windows 10, Windows 11, and Ubuntu Linux 21.04 without platform-specific difficulties.

It is extremely helpful, although not essential, to have a multi-monitor setup. That way you can run the demonstration in one screen, and the Teams call on the other.

You should also, as discussed above, be fairly comfortable with R basics before you enrol on this training. Do please get in touch with brendan.clarke2@nhs.scot to discuss if you are unsure.

Joining instructions

You’ll need to do a little bit of preparation before the first training session, which should take about 15 minutes to complete. Please make sure you have completed this before the start of the first session so that we can make a prompt start. If you’re new to RStudio Cloud, please follow the step-by-step instructions below. If you’ve worked with RStudio Cloud before, you can just sign-in to your account at RStudio Cloud, create a new project from the GitHub Repository at https://github.com/bclarke-nes/Intermediate-R.

It is also possible to complete the training using the desktop version of Rstudio. This training was written using R 4.2.1, and you may encounter some difficulties related to versions if your installed R is very different from this.

Step-by-step instructions

1. Go to https://rstudio.cloud/
2. If you have an account, you can log in as normal. Otherwise, please create a new account by selecting Get started for free, following the steps, and then signing-in
3. Once you’ve signed-in to RStudio Cloud, add a new project by clicking New Project >> New Project from a Git repository. When prompted, enter the URL https://github.com/bclarke-nes/Intermediate-R
4. That will give you a new project containing the files needed for this training

Aims

• To introduce a series of R/tidyverse tools that are particularly helpful to analysts moving beyond R basics
• To provide a social learning space to support learners as they develop their skills into these more challenging areas
• To show how more advanced R techniques can be used in real-world data analysis in health and social care

Objectives

By the end of these session, the user should:

• Have developed their understanding of functions, iteration, dplyr, purrr, and tidyevaluation
• Have connected these new functions to their own work
• Have participated in social learning during the training sessions, reviewing the code of other trainees
• Have confidence in reading third-party code that uses some of these advanced R techniques