Last week we were entertained by a professor from the States. Jared Lander (@jaredlander) joined us for a project we currently are involved in. Apart from being a really cool guy, he also gave some ideas of what a DataScience Tint for the hex could look like.
To give you some context, the first thing you need to know about Jared is that he is the author of “R for everyone” (get the book, it is really awesome). So we suspected that the first recommendation he would make for the tint would be to install R onto it. Needless to say that we put that one on top of the list.
Since nor Wim nor I are data scientists (allthough Wim is making up for that in a burning pace) we did have some questions on what we could actually put into a tint like this. The whole concept of a hex is to provide everything needed to get the job done. This does mean we also put tooling into a tint. Jared pointed us to RStudio Server for that one. What can I say, I like nice web interfaces. Second on the list that is.
But R Studio would run on a single node, so how could we use the whole storage/processing capabilities of the hex? There is no direct need to install hadoop for doing this, so we went out and looked for alternative distributed systems. We are currently looking at Ceph for the storage while we still have no concensus on the processing part. Maybe install different types of data stores on different nodes.
Anyway, it was a great (but heavy) week. And since Jared is the owner now of a nice blue hex, I’m sure we can expect a lot more feedback from him in the future.