Skip to content
 

Plans for reboot of Statistical Communication class

At the end of my course on Statistical Communication and Graphics last semester, I enlisted some of the students to help plan for the new version of the course (which starts next week). I took a bunch of notes on the blackboard and then a student took pictures for me. I had the idea that I’d use these notes to plan the revised course. The discussion was helpful, and it was probably even helpful to write all this stuff on the board, but I didn’t really know what to do with the notes themselves . . . so I’m sharing them here below, just in case they amuse you. Perhaps a good message to send to all of you, that I don’t know what I’m doing either:

photo 1

photo 2

photo 3

5 Comments

  1. Bob says:

    Have you ever thought of using this textbook for your course?

    Introduction to Statistical Communication Theory (Middleton, first ed 1960) You would probably like it. The blurb states “Using the statistical Bayesian viewpoint, renowned author David Middleton employs . . .”
    http://www.wiley.com/WileyCDA/WileyTitle/productCd-0780311787.html

    Another classic is http://www.amazon.com/Introduction-Theory-Random-Signals-Noise/dp/0879422351 Not so Bayesian as I recall, however. But, a classic (1958) in the field.

    Or, are you trying to teach about communicating statistics rather than statistics in communication engineering?

    Bob

  2. So, which refs do you want? I was recommending Hunt and Thomas’s The Pragmatic Programmer as a good general overview of how to set up projects and workflow. It’s more aimed at long-running projects and professional coding than at one-off stats experiments or projects, but it’s a good read.

    When I guest lectured in your class last semester, I broke it down into three bits, really:

    1. You should be using version control. This is also discussed in The Pragmatic Programmer, and I see GitHub as an item on your notes. Teach yourself how to use it, first! We don’t need to get into complications, just something so that we can all check out the same thing, work on it, and merge to a shared repo and look at diffs.

    2. What is floating point? I went over how floating point numbers are represented in IEEE arithmetic, with stats-relevant items such as why you get underflow in densities not on the log scale, why you can have numbers very close to 0 (such as 1e-300), but you can’t get numbers very close to 1 (closest you can get is around 1+1e-14 or so), why you can’t take an exp of a value less than -1000 or greater than 1000 without overflow or underflow, and why subtraction can be so nasty. I also talked about NaN and +/- infinity values. The Wikipedia pages on floating point arithmetic are quite good.

    3. Designing an API. I went over what you had to do to fully specify an application programmer interface, with some simple stats examples that built from sum to mean to sd, talking about boundary conditions, error handling, and doc, and why these are all so interrelated. There aren’t really great references I know for this, either, that aren’t language specific. Like Bloch’s Effective Java, Norvig’s book on Lisp coding for AI (ancient, but a classic), Meyers and Sutter on C++, etc. I don’t know good refs here that relate to R or Python, but would love to hear about some so I could read them myself.

  3. Keith O'Rourke says:

    From what I can make out from what’s on the boards and your last statement.

    Maybe – Questions, answers and statistics, 1986 by Terry Speed http://iase-web.org/documents/papers/icots2/Speed.pdf
    (I just came across it and was impressed.)

    And to complete good/bad with the bad and ugly – Criminals in the Citadel and Deceit all along the Watchtower: Irresponsibility, Fraud, and Complicity in the Search for Scientific Truth Prathap Tharyan

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3353596/?report=printable

  4. Elin says:

    One question is, who are they supposed to be communicating to? The public or writers who will use their graphs or results? One way to see how writers are told to think about numbers is http://www.press.uchicago.edu/ucp/books/book/chicago/C/bo3636131.html though there are definitely parts that drive me crazy. It’s definitely oriented towards writers who want to incorporate numerical information, but for statistics students it might be a good guide to writing. Also there are a few places that she gets it wrong, but those are really interesting to talk about. If you are trying to communicate it is good to really consider where do people have the most problems, which is sometimes about language (see the title of this for a good reason to parse titles before publishing http://www.southerneducation.org/Publications/New-Diverse-Majority/Percent-of-Low-Income-Students-in-PS-2015-01.aspx ).

Leave a Reply