Emma Rand and Mine Çetinkaya-Rundel

In September, we will be delivering a series of four online one-hour package development workshops. These are:

About R Forwards

R Forwards is an R Foundation task force that was originally set up in December 2015 to address the underrepresentation of women in the R community. An analysis of CRAN package maintainers estimated that fewer than 15% were women, and a useR! participant survey found that women were less likely than men to have experience contributing to or writing packages. In 2017 it was rebranded to accommodate other under-represented groups such as LGBTQI, minority ethnic groups, and people with disabilities.

Why create Package Development modules?

Forwards have delivered face-to-face one-day workshops in Package development, supported by a grant from the R Consortium, for several years. These are heavily based on the R packages book by Hadley Wickham and Jenny Bryan. These have worked well but our reach is limited because only those with access to teaching facilities can deliver the teaching and only those able to attend a face-to-face workshop can benefit from it. This misses a lot of people! Recently, we have been modularising our materials into one hour long workshops that can be delivered online or in person. We hope this makes the material more usable to teachers and learners alike.

Some module design principles

By December we had begun work on our module design principles. Just like the face-to-face-workshops, the modules would teach package development using devtools in RStudio. Our aim is to provide workflows to help people get started rather than an exhaustive understanding of the details (you can see Writing R Extensions for that).

We wanted the collection of modules to be relatively short (~1 hr), ‘stackable’ and easy for others to use. We thought each module should:

• be approximately 1hr
• be discrete (standalone) but link to other modules
• have specified prerequisites and learning objectives
• be a complete resource for teaching (a person should be able to teach themselves from the material) and include tutor notes
• have a set of Rmd slides with comprehensive alt text and speaker notes
• have slides with minimal content but detailed speaker notes
• use live coding, (minimise ‘lecturing’); include additional exercises for the speedy

Our Progress

In January, Mine and Emma set about developing a module template and the first three modules. We delivered these over consecutive days in February (15th, 16th and 17th) to a great bunch of people. A total of 51 different people took the modules. We were expecting most to do all three but only 19 chose to do all three and 21 people did just one. Perhaps a modular approach did help people tailor training to their needs and availability?

We’re delighted that at least one of our participants, Melissa Wong, has now released a package on CRAN! The package, pomcheckr, implements the method described at UCLA Statistical Consulting for checking if the proportional odds assumption holds for a cumulative logit model.

If you’ve been to a Forwards Package Development workshop and have released a package we would love to hear from you.

If you’d like to go to a Forwards Package Development workshop and have release a package Register at the top of the page!