Celebration2020 – A great get together to celebrate 20 years of R-1.0.0

Last week (Feb 28th – Feb 29th) the celebration of 20 years R 1.0.0 took place in Copenhagen, Denmark. It turned out to be a fun conference. People told lots of anecdotes from the earlier days of R. Peter Dalgaard brought with him the very first CD with R-1.0.0.

Visitors had the possibility to meet and see more members of the R-Core team. Hadley Wickham’s spontaneous talk generated a lot of laughter. The location in the Maersk tower allowed a fantastic view over Copenhagen. I really enjoyed this well-organized conference. In the next few sections I’ll go through my highlights:

ggplot workshop from Thomas Lin Petersen

The workshop was really well structured. Thomas started by explaining the grammar of graphics that I had actually never heard of. It’s the basic idea behind ggplot. The naming conventions of all functions follow these principles. The grammar of graphic implies the order of function calls for a ggplot, too. You can read on it in this book or look for Thomas presentation. He hosted the workshop in RStudio cloud. It contained little tasks to solve and explanations of the basic ideas. You can find the project here, if you want to go through the workshop tasks on your own.

Peter Dalgaard on the history and future of R

Peter introduced the whole story of R. It was interesting to hear that until 1997 the whole developer team only worked remotely. There were people developing the R-core in Auckland (NZ), Vienna (Austria), Graz (Austria), Copenhagen (Denmark) and many more places. They all first met in Vienna in 1997 in person. Then decided to release R on the nerdiest date possible. The 29th of February 2000 is a date, that has three exceptions when to introduce a 29th in February (read more here). The first release of R did not only contain the R-core, but also the whole CRAN at this time burned onto a CD. This CD got copied 15 times and signed by all Core Team members. In Auckland it even has its own shrine.
The highlight of this presentation was Peter releasing R-3.6.3 live in front of the audience. He uploaded it to CRAN and changed the CRAN website content. Here is a time-lapse video of the live-release:

Roger Bivand: How R Helped Provide Tools for Spatial Data Analysis

Roger showed in his presentation how he started in the year 2000 to do geographical data analysis. I was really surprised that all these packages exist for so long. In several projects, I used the packages to map sports data to maps. It was great to see that the API has changed a lot, but his team made these things work for over 20 years now. I highly recommend looking into the sf package now, if you want to start with such things.

Mark Edmondson – Google Cloud

For me Mark’s talk was a real wow talk. In my open-source projects, I always want to allow other users to reproduce my code, or see how my package was checked. Thus, I used the Docker-based tool Travis-ci. Mark is working out packages for the Google Cloud. He showed that the Google cloud can easily allow you to trigger and run build processes. But the Google Cloud also offers to run such processes on nearly any kind of event. He presented a tool that triggers a plumbr API call upon just uploading a file to Google Cloud Storage. This was a really impressive piece of software. As you can imagine updating any kind of dashboard based on such events. His package googleCloudrunner also has a very nice API to fulfill such tasks. Wonderful.

Julia Silge – Text Mining with Tidy Data Principles

Julia is really great at giving entertaining talks. Her talk was full of fun facts of what you can do with text-mining. Her great work included showing what movie programmers identify with and which states are overrepresented in American song lyrics. Her package tidytext allows an easy and really flexible analysis of any kind of text data. Julia is a big Jane Austin fan. Her analysis of books are amazing. She made it possible to predict which book a phrase comes from. This was possible only on training an algorithm on different Jane Austin books. My favorite algorithm was the topic finder. It could easily find in reddit analysis, which words belong to certain topics. This would allow for automatic assignment of topics for articles. You can find such work explained here.

Hadley Wickham – Why R is a weird language.

Winston Chang got sick and that is why Hadley Wickham jumped in to give a remote talk. He went over three aspects, why R is weird. Functions can be stored nearly anywhere and you can make everything a function. You can even make a function return a function. Not too crazy. Environments really allowed him to play a lot. He actually overwrote the + operator inside an environment, that he used to execute a function. You can even change the environment of just this single function called add.

Expressions are a cool way to store calls. I use them to do reproducible code. But you can also use them together with environments to run the same call with different inputs.
Hadley’s whole talk was live-coding. It was impressive how he can make people laugh because he puts in great hacks. There were fantastic things he used to overwrite even the "(" inside R. The code from this talk can definitely be used to mess with your co-workers. There will be some people in trouble in the next weeks.

Heather Turner- R Forwards

Heather grabbed an important aspect from Peter’s talk. The R-Core consists 1) of old people 2) of men 3) of people from highly developed countries. R Forwards is working on the integration of more diverse people into the R community. This not only includes projects like R-Ladies and AfricaR, but also projects to enable blind and deaf people to use R. Her talk presented in which direction I would like to see R going. It should be a more democratic, more transparent language. It was impressive that she used automatic captions during her whole talk, allowing deaf people to follow. One important point of her talk was opening up the R-core for more people. Right now it is pretty hard to collaborate on the R-core. Python went over this problem by several steps. Additionally, the R Forwards project will enable more ways to learn R and to ask questions. Stackoverflow, github, and CRAN do not contain content that helps total beginners. R Forwards wants to change that by workshops, setting up easier guides to work with R, creating a kind of mentoring program, finding a tool where people can ask any kind of questions. Heather presented brilliant ideas. Check them out at: https://forwards.github.io/

This article is just an excerpt of what happened around the conference. So please also read on the topics I missed out:

• Hannah Frick: Assessing model parameter stability in R
• Therese Graversen: R Provides Statistical Evidence in Criminal Case
• Dirk Eddelbuettel – Rcpp Workshop: Extending R with C++