useR! 2018 Confernce: Brisbane 10-13 July 2018.
I had 2 main objectives in attending this year’s global “useR! 2018 Conference”, the global R users’ conference in Brisbane:
1. To develop R programming skills;
2. To reinvigorate my passion for building data science solutions using R.
R “hex” map of Australia, with each hexagon representing an R package
Develop R Programming Skills
I sat in on presentations about some cool and useful packages in many of the sessions, but all that they did was point me in the right direction and then leave me complete the journey on my own. As is the case for all technical skills, there’s no magic pill and no substitute for rolling up the sleeves, reading the documentation and learning through trial and error.
Reinvigorate my passion for building applications in R
Right up until the final presentation, I felt that the sessions were more-or-less interesting but the conference was lacking any “come to Jesus” moment. That said, it doesn’t help that that I was feeling anxious about limited opportunities I get to practice R day-to-day. On the plus side, the conference reminded me of the huge potential in R, so long as I continue to work on my R fluency.
Conference Format
- Tuesday all-day and Wednesday morning were devoted to optional half-day tutorials, which I purchased.
- Wednesday afternoon and all-day Thursday and Friday saw a smorgasboard of presentations:
- Keynote speakers, mostly from Australia but some international R-world celebrities.
- Half-day threads of 20-minute presentations (such as “Applications in Society” or “Big Data”). Many of these were from academics or PhD students but plenty were from business speakers.
- Lightening talks and poster presentations.
- Cocktail drinks on Wednesday evening.
- A formal dinner on Thursday evening.
Favourite Presentation
My favourite presentation would be Code Smells and Feels by Jenny Bryan, the final presentation of the conference. Jenny’s very engaging style worked well as she described tips and tricks to writing elegant and decipherable code. Slides for her presentation are here: https://speakerdeck.com/jennybc/code-smells-and-feels. It helped that I read an interview with Jenny earlier this year and have developed a bit of a geek-crush.
Honourable mentions:
Just about all of the presentations I attended, but in particular.
The Ultimate Online Collect Toolbox by Hanjo Odendaal. Hanjo went through the basics of scraping static web pages using the rvest package and then the intricacies of navigating dynamic pages using Sellenium. I’m still to get my head around using Sellenium and Docker but the introduction and overview was exciting.
Interactive Data Visualisation on the Web with R by Carson Sievert. Carson is developing a new package for writing Javascript objects to liven up web presentations based off R applications. Carson also brought to my attention how I can easily add interactivity to my own Shiny applications.
The workflowr R package by John Blishchak. I sense that there are a few workflow packages emerging for R but John’s looks as useful and any I’ve seen.
The Grammar of Animation by Thomas Lin Pedersen, which presented a philosophy for understanding the strengths of and differences between animation, graphics and interactivity. Any presenter who channels Hans Rosling’s Gap Minder animation gets my attention.
Using R to help industry clients - the benefits and opportunities by Lisa Chen. Lisa is a friend of mine from New Zealand and does a wonderful job in selling me the strengths of R. A retrospective of the growth of R among major organisations in New Zealand was exciting.
Teaching R to New Users: From tapply to Tidyverse by Roger Peng. Roger, from John Hopkins University in Baltimore, delivered my first Coursera course back in 2013, so it was super exciting to hear him speak live. Roger’s retrospective of the evolution of R was interesting, though the comments afterwards from the hard core R programmers was enlightening.
Tidy forecasting in R by Rob Hyndman. Rob presented on the next major upgrade to his forecasting package, which adopts the “tidyverse” approach and has so many new features that it required a a completely new package, “fable”. I’ve been star-struck by Rob ever since meeting him at an international forecasting conference in 2013, amazed by his self-assurance and his contribution to the science of forecasting time-series.
Graciously, the conference organisers videoed all of the presentations for viewing on the R Consortium channel on YouTube.
Disappointments
The biggest disappointment was the apparent lack talks on how to scale up analysis of data sets greater than 2 million records. I’m sure that there were some presentations but I’ll need to continue searching.
A lesser disappointment was the awkward nature of many of the presenters. It was Hadley Wickham, one of the world leaders in R, who put me onto a Coursera course in public speaking and the keynote speech. It was a shame that some of the techniques introduced in this course and others like it were hardly seen at the conference.
Final Note Thank you so much to Di Cook and her team at Monash University for successfully campaigning for Australia to host a useR conference and then for organising such a great conference. Although I had to take the week as annual leave and pay my own dues, it was well worth it: I’m reinspired to continue my R development as well receiving great suggestions for where to focus my next learning endeavours.
Your’s truly in the conference foyer
Lisa and Phil from Harmonic (NZ)
Conference dinner on Thursday evening