My Channel 9 interview from Brisbane

Here’s an interview I did with Microsoft Channel 9 right after my talk in Brisbane.

You can find the interview in multiple audio and video formats on Channel 9.

9 thoughts on “My Channel 9 interview from Brisbane

  1. Right, the prohibition on using underscore as a variable name word separator is no longer a part of the language, but it persists by convention. So does the use of a dot as a word separator that arose as a compensation. Just backward to the majority of the software world.

  2. Nice interview, though I’ve asked more specific questions if I had the chance. Given that you’re working a lot with experimental data I’m curious about the data management process. I guess you have some infrastructure for data acquisition -> storage -> selection -> analysis. I expect R to come into play in the last two steps, although I’ve seen people heavily relying on matlab’s mat files all the way through the whole process. Could you elaborate on non-secret parts of your workflow? Things like how you get data into R’s workspace and what happens before that. Just got curious listening to some of your interviews on ch9.

  3. Most of the R code I deal with is generating its own data via simulation, so data management isn’t an issue.

    But some of my colleagues who do manage data and process it with R use Sweave to keep data analysis in sync with the data. See my previous blog post on Sweave.

  4. You look much younger at the video than at the photo in the blog.
    You either should replace the photo or the video.

  5. John,

    Nice interview. I am curious about the C# numerical library that you mentioned. Is that library available – either as open source or commercially? Have you seen the NMath library? If so, what do you think about it? I understand it uses Intel MKL underneath.


  6. I looked at several C# numerical libraries, and none had what we needed. The library is not available publicly, but it probably wouldn’t be generally useful: it has grown to meet our peculiar needs. It has things you won’t find elsewhere, and lacks many things that a general library should have.

  7. I didn’t know you were at MD Anderson. I’m a medical physicist, and of course MDACC is extremely distinguished in the field (I studied at LSU under Ken Hogstrom, who was the director of Medical Physics at MDACC for several years). Guess I should take time to get to know a little about the bloggers I subscribe to. I don’t remember what post I read that led me to subscribe, but I remember liking your style. It’s interesting how we are drawn to people we have things in common with. Thanks for the always-interesting posts!

  8. Interesting interview. I used R extensively a few years back and before that S-Plus. You’re right, very good for simulation and makes numerical statistics easy.

    No, it just doesn’t scale, as you point out. Thus, it can be hard when using with external data generated by say, a Google API. I think a major challenge data science now faces is getting R-like simplicity for big data.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>