Last week I took part in the BaselR meeting hosted in the very nice Roche Learning Center Auditorium and organised by Mango-Solutions. I find these informal meetings mostly organised via meetup perfectly fit the nature of what data science seems to be currently: it’s organic, there is still not defined frame, things are moving very fast and people have to meet.

There were three talks summarised below.


FFTrees was presented by its creator, Nathaniel D. Phillips. This package aims to create fast and frugal decision trees. That’s fit perfectly in my needs: be pragmatic with statistics. Furthermore the visualisation produced is very informative. I already have some ideas for implementations in my field.

Reproducible analysis

I presented a pragmatic reproducible analysis workflow with R. If you are following my blog, you know that’s a topic I’m working on. That’s why I gave another talk on this subject, showing how I extended my workflow with some solution to archive a part of software layer thanks to packrat (more about this in a later post on this blog).

R with Python

Rita Giordano talked about Structural biology with R. I had a strong interest in the way she shows us how to use R with Python with Rpy2. The point was not the traditional R VS Python debate but instead how diversity of these programming languages can lead to powerful data analysis. By the way, in our next R user group in Strasbourg we will be talking about R vs Python!