20
Aug

A couple of requests for the @Statistics2013 future of statistics workshop

Tweet about this on Twitter8Share on Facebook4Share on Google+0Share on LinkedIn1Email this to someone

Statistics 2013 is hosting a workshop on the future of statistics. Given the timing and the increasing popularity of our discipline I think its a great idea to showcase the future of our field.

I just have two requests:

  1. Please invite more junior people to speak who are doing cutting edge work that will define the future of our field.
  2. Please focus the discussion on some of the real and very urgent issues facing our field.

Regarding #1  the list of speakers appears to be only very senior people. I wish there were more junior speakers because: (1) the future of statistics will be defined by people who are just starting their careers now and (2) there are some awesome super stars who are making huge advances in, among other things, the theory of machine learning, high-throughput data analysis, data visualization, and software creation. I think including at least one person under 40 on the speaking list would bring some fresh energy.*

Regarding #2 I think there are a few issues that are incredibly important for our field as we move forward. I hope that the discussion will cover some of these:

  1. Problem first not solution backward. It would be awesome if there was a whole panel filled with people from industry/applied statistics talking about the major problems where statisticians are needed and how we can train statisticians to tackle those problems. In particular it would be cool to see discussion of: (1) should we remove some math and add some software development to our curriculum?, (2) should we rebalance our curriculum to include more machine learning?, (3) should we require all students to do rotations in scientific or business internships?, (4) should we make presentation skills a high priority skill along with the required courses in math stats/applied stats?
  2. Straight up embracing online education. We are teaching MOOCs here at Simply Stats. But that is only one way to embrace online education. What about online tutorials on Github. Or how about making educational videos for software packages?
  3. Good software is now the most important contribution of statisticians. The most glaring absence from the list of speakers and panels is that there is no discussion of software! I have gone so far as to say if you (or someone else) aren't writing software for your methods, they don't really exist. We need to have a serious discussion as a field about how the future of version control, reproducibility, data sharing, etc. are going to work. This seems like the perfect venue.
  4. How we can forge better partnerships with industry and other data generators? Facebook, Google, Bitly, Twitter, Fitbit etc. are all collecting huge amounts of data. But there is no data sharing protocol like there was for genomics. Similarly, much of the imaging data in the world is tied up in academic and medical institutes. Fresh statistical eyes can't be placed on these problems until the data are available in easily accessible, analyzable formats. How can we forge partnerships that make the data more valuable to the companies/institutes creating them and add immense value to young statisticians?

These next two are primarily targeted at academics:

  1. How we can speed up our publication process? For academic statisticians this is a killer and major problem. I regularly wait 3-5 months for papers to be reviewed for the first time at the fastest stat journals. Some people still wait years. By then, the highest impact applied problems have moved on with better technology, newer methodology etc.
  2. How we can make our promotion process/awards process more balanced between theoretical and applied contributions? I think both are very important, but right now, on balance, papers in JASA are much more highly rated than Bioconductor packages with 10,000+ users. Both are hard work, both represent important contributions and both should be given strong weight (for example in rating ASA Fellows).

Anyway, I hope the conference is a huge success. I was pumped to see all the chatter on Twitter when Nate Silver spoke at JSM. That was a huge win for the organizers of the event. I am really hopeful that with the important efforts of the organizers of these big events that we will see a continued trend toward a bigger and bigger impact of statistics.

* Rafa is invited, but he's over 40 :-).**

** Rafa told me to mention he's barely over 40.