useR! 2024

Data science is an ever-evolving industry that requires constant upskilling. The pressures to learn the latest tools for project deliverables or to enhance one’s CV can be a hindrance to effective learning. Here, I argue for the need for silliness when developing new R skills. Learning is far more enjoyable and conducive to retention and application when we take away the seriousness of upskilling. I will share my experience in creating {ohwhaley} - a ‘toy’ R package which serves as a tool for learning package development and upskilling new learners. I hope attendees will walk away feeling more light-hearted and empowered to build silly things in R to reinvigorate their curiosity for R knowledge.

We present a method to teach hypothesis testing by engaging students into an A/B-test facilitated by a Shiny app.

The idea is to teach hypothesis testing in the context of A/B-testing, which is nowadays massively used to optimize apps and webpages. The proposed method engages students into an A/B-test in class as follows: Students access a quiz app via a QR-code. The students are randomly assigned to either control or treatment group and run the quiz without knowing their group or the treatment. We then reveal the treatment and anonymously save and analyze the scores in class using a hypothesis test to see whether the treatment had any effect on the scores.

The procedure, based on a pen&paper game by Adam Shrager, comes with a number of challenges and pitfalls – as any randomized trial - that can be openly discussed with students.

In this presentation we walk you through this A/B-test, collecting live quiz data from the audience if time permits. We present and discuss some challenges and key findings. The app is freely available, feedback is appreciated:
documentation and installation: https://github.com/IFAFMultiLA/memory_game
live demo: https://tinyurl.com/MemGameTrial

Clinical research is one of the fastest growing fields in the world, and R is becoming increasingly important as a way to handle data, especially as more and more studies are conducted with small numbers of patients, or in collaboration with multiple institutions to collect data and conduct research. Rather than using R to analyze data, clinical researchers have typically focused on study design, data collection, and validation, while coding has been done by professional developers, but now more and more clinical researchers are trying to use R themselves, including data management. To this end, we have been providing R training for clinical researchers, but there is a lot of room for improvement compared to professional training services, such as reflecting the latest R-related technology trends and making the training experience better. In this session, I will share how we decided to use Quarto, what we considered in order to provide R training for clinical researchers, how we actually used Quarto, the advantages and disadvantages of using Quarto, our achievements, and our future plans.