Blog / Opinion

Latest news, events and opinions

Sawtooth/SKIM training on R

Last week, I had the opportunity to join Sawtooth and their European partner SKIM on their training in Rome. As topic of interest, I joined a full day training on R. Next to some training sessions, there was also the conference, which I attended as well.

 

At the conference there were 10 presentations of different MR agencies on various topics, mostly related to choice modelling and segmentation.

 

As you can imagine, there are definitely worse places in the world to follow a training than the sun-bathed city of Rome. However, we were not there for fun, but for training, so I spent most of the day sweating over R code in a meeting room.  

So why R… First off all, R is a collaborative programming language, specifically designed for statistical analyses. Thousands of people have contributed to the package of R, which in itself is free to use, making it one of the largest statistical software’s out there. Secondly, as everyone can contribute to the software, it makes it one of the first places where new ideas and new applications can be found. This makes this program very valuable and powerful. 

But there are always drawbacks:

The biggest drawback is that it is a scripting language and not a software. Unlike other software programs, it has no graphical user interface, nothing that explains you what you can do, or what different options you have. Secondly, R itself knows almost nothing. It is an empty shell where you can install or create packages that run the analyses. In total there are now over 8000 packages already, making it hard to find just that one package you need, and each having their specific code and syntax. 

So what is the best way to start learning R?

Somehow I was reminded at my first steps of learning French, now 22 years ago. We used the book “Eventail” and the first few phrases were “je suis Jerôme, je suis très fort.”  And though they seem very phrases when I think about them now, they were the ones on which loads of other vocabulary and grammar were built on. And through repeating them over and again, they were beaten in.

After the training, I even more strongly believe this is how you need to learn R. Through basic small steps, doing part of the analysis in R to try the software and code and to learn the output and usability. So that one day, many years later, you find yourself in France, and someone steps up to you, asking for directions. And somehow, miraculously, you know what to say to them to put them on the right track. 

And one further important thing I learned at the training, you’re not alone on your road to R. Often your questions have been asked before, so try searching for them. Also R studio is a nice piece of software. It is a program designed to use R, where you can separate your code from the output making rerunning analyses and saving parts of the analyses easier. And lastly, practice, practice and practice are the most important things to remember. 

by Frederik De Boeck,
Senior Data Scientist
on 19-04-2016

Ready to get inspired about new thinking & tools in data analytics & visualisation #DSBXL27 https://t.co/gtCsF0J91d

@boobook_world