The Fun of Building Things and the Challenge of Learning – the rOpenSci OzUnconf 2019

It was the best of times, it was the worst of times.

Dickens might have meant it figuratively, but in the case of the rOpenSci OzUnconf 2019, we mean it literally. Set to the backdrop of a national emergency that is still ongoing from 11-13 December, our participants came from across Australia as well as New Zealand, Japan, India and Indonesia.

An rOpenSci unconference (unconf) is about building – building software, tools, but more importantly community. It operates on the premise that we all belong.

Joining the R community doesn’t start with an entrance exam – there is no point at which you’re ‘good enough’ to belong. A person belongs the moment they decide to join in – sometimes people join in after a few decades of programming, sometimes they join in before they open R for the first time. An rOpenSci unconf is intended to support that community and provide a place to bring a diverse group together in order to learn from each other and build things.

An unconf cohort is diverse by design – not just gender diverse (though this is important), but diversity of skill, background and interest. Most participants do not know each other – and this is another feature that is deliberate. An unconf doesn’t start as a group of friends building things together – but it is often the way it ends.

Our initial pre-conference conversations included the potential projects we wanted to work on as well as whether or not participants with respiratory illness may require protective masks, if they could be found.

Saras Windecker and Nick Tierney gave interested participants a git 101 workshop on building R packages and using git and GitHub on the day before the unconference itself. One of the purposes of this workshop is to ensure everybody is ready and confident to jump in from the very beginning.

In the midst of the chaos of a climate and health emergency for many Australians, the unconf team found something to do about it – and project Smoky was born. The project aimed to provide a way for the community to access relevant and timely information on air quality, which was difficult to source.

During the unconf, we were working in the maths and statistics department at the University of Sydney, and we had a number of participants who know how to control chaos and use it for their own ends. To whit, some team members decided to generate valid R code randomly using BNF grammars (for those of us who don’t automatically know what those are – I think of it as ‘grokking the rules that make R work’). Others held an impromptu aRt class and exhibition, generating art entirely within R itself.

Other teams set out to counter the chaos sometimes present in an operational data science environment with actions_sandbox, which explores how GitHub actions can automate useful tools for improving productivity. Other groups created a tool – Rcademy – for managing academic promotion, or made gghdr, a ggplot extension for plotting highest density regions!

There was a selection of data-based projects at the unconf this year, with one team, OZdatasets, working on expanding and maintaining the Australian data sets collected at previous OzUnconfs as well as other interesting or unique Australian datasets and resources. Another team brought together international education data to unpack some misperceptions around Australia’s rankings in learningtower. The project quizlite provides infrastructure to create light-weight quizzes for use during teaching (and other places) and in the spirit of the unconf, one team member built the praiseme package (because who doesn’t need a little support while writing R code?) into a web app (because that need isn’t R-specific).

Check out all our projects and contribute if you’d like: remember that there’s no entrance exam and you’re part of the community from the first time you decide to join in – online or in person. If you’re not sure where to start, but would like to get involved, start by opening an issue in a repository and ask where you can help- team members will be more than willing to support you, the way they’ve supported each other.

Jenny Richmond was the WeAreRLadies Rotating Curator during the unconf. She generously drew attention to all the project teams and the many R-Ladies chapters that were represented.

Team 3 @rOpenSci #ozunconf19 were working on PISA data. There has been some media coverage recently about how Australian students math scores are declining. This group are pulling the PISA data into an R package so that more people can analyse it. pic.twitter.com/cGM9zW7FQM

— We are R-Ladies (@WeAreRLadies) December 12, 2019

Our OzUnconf was made possible by rOpenSci, who originated this style of hackathon-flavoured unconference in 2014. The rOpenSci team were incredibly generous with their time, experience and patience.

The generosity of our sponsors provided us with the means to hold the event and make participation possible for some who might not otherwise have been able to attend – notably the United States Studies Centre and the Centre for Translational Data Science at the University of Sydney. We also received support from Nous, the Australian Bureau of Statistics, Internetrix and the Statistical Society of Australia.

Special thanks also go to John Ormerod, Emi Tanaka, Rob Hyndman, Di Cook, and Nick Tierney – without whom we would have had maximal chaos and no control at all! Unconfs build on each other, and previous unconf teams’ work helped us successfully hold our own event. We’d especially like to thank Nick Tierney, for providing guidance in transferring the leadership for this OzUnconf, passing on his knowledge of being the lead organiser of the original three OzUnconfs. We’d also like to thank Miles McBain and Jesse Roberts, who, along with Nick, helped create the first rOpenSci OzUnconf back in 2016. It has been a joy to see this community develop and thrive.

If you’re interested in joining the community, remember there’s no entrance exam – you’re welcome whenever you choose to join in!