Talking at LMA in Agrocampus Rennes
Posted on November 30, 2013
On November 29 (2013) I had the nice opportunity to visit my French colleagues at the LMA2: Laboratoire de Mathématiques Appliquées (Applied Mathematic Lab) in Agrocampus, Rennes.
The perfect excuse was Sébastien Lê’s invitation to give a talk to the grad students specializing in applied statistics.
In case you are wondering who’s part of the LMA or what they do? Well it happens that they are one of the most active French groups working with R, and creators of some really cool packages for data analysis like FactoMineR, SensoMineR, missMDA, and FAMT.
This time I was kindly hosted by Julie Josse, who is always enthusiastic to talk about her research and other topics related with data analysis. I also took advantage of my visit to keep the old book signing tradition and ask Jérôme Pagès to sign me a copy of his most recent book (in French) Analyse factorielle multiple avec R.
Context of my talk
You can check the slides of my talk here
Sébastien asked me to make a presentation for the grad students on the work that I’ve done. First I thought about presenting some case study or some specific statistical analysis application. But then I decided to focus on something slightly different. Instead of giving the typical talk about an XYZ method and how to apply it, I chose to provide a synthesis of my programming experience. More specifically, on my experience developing stats software from within my academia research career.
Basically I’m sharing some personal thoughts around developing stats software at a meta level framework. I didn’t talk about any technicalities or how to program. What I presented has to do with the general implications of my work. The take-home question is very simple: What it is that we achieve with a computational-analytical-tool? Personally, when I develop a package, I would like it to help me address at least one of the following principles: