Skip to content
 

Attractive models (and data) wanted for statistical art show.

I have agreed to do a local art exhibition in February.

An excuse to think about form, colour and style for plotting almost individual observation likelihoods – while invoking the artists privilege of refusing to give interpretations of their own work.

In order to make it possibly less dry I’ll try to use intuitive suggestive captions like in this example TheTyranyof13.pdf

thereby side stepping the technical discussions like here RadfordNealBlog

Suggested models and data sets (or even submissions) would be most appreciated.

I likely be sticking to realism i.e. plots that represent ‘statistical reality’ faithfully.

K?

6 Comments

  1. Phil says:

    I can't figure out what TheTyranyof13.pdf is displaying, can't interpret the caption, and can't understand this post. I'm not sure if it's just me being dense; if so, I apologize.

  2. JohnnyZoom says:

    I, too was flummoxed by that pdf. But what really got me going for a few seconds was the title of the post.

    Let's just say I interpreted "attractive models" in a different way than was intended :)

  3. K? O'Rourke says:

    Phil: I am trying to have some fun – but not at anyone's expense …

    For instance JonnyZoom got a smile out of the "attractive" miscueing in the title I hoped would work that way.

    It is meant to be an art show – with the accepted artist's privilege of refusing to give interpretations and even worrying about how widely their work can be interpreted.

    The TheTyranyof13.pdf might only make sense to Radford – and that should be OK. (There is full technical discussion of what I am trying to represent in the plot on his web site – if others want to give it a go.)

    Now, I will likley entitle the show "Painting Statistical Inferences: The realism of relatively surprising if not pretty pictures."

    For anyone to even have a glimmer of what the "relatively surprising" phrase is alluding to they would need to familar with this (very technical material) http://www.utstat.utoronto.ca/mikevans/research.h

    But all and all, I am trying to evade the tyrany of being forced to explain things I grasp in terms others will easily grasp. That is sometimes my work, but this is going to be my art show ;-)

    K?

  4. Phil says:

    K?, Even an artist with two eyes sometimes enjoys grilled fish, if you see what I'm saying. Not trying to rattle any cages, just saying it like it is.

    I think I get what you're looking for, now that you've explained it: colorful results (in both senses, ha ha!) of both likelihoods and probabilities from a surprising model (in both senses, ha ha ha!) that is neither tyrannical nor explainable, or, if explainable, doesn't need to be explained to be colorful and surprising. Correct me if I'm wrong.

    As you can probably guess, I still haven't a clue what you are looking for. Which is OK with me if it's OK with you.

  5. K? O'Rourke says:

    Phil, yes hopefully "doesn't need to be explained to be colorful and surprising"

    So looking for data sets and models that would give "pretty pictures".

    John Nelder's "There are no outliers in ther stack loss data set" will likely work – extreme outliers under usual assumptions become well in line with other observations under (arguably better) assumptions.

    Anscombe's Quartet of "Identical" Simple Linear Regressions (see help in R) may also work into something artistic.

    Probably anything that seems surprising in a data set _given_ a particular model could be worked into something.

    When I get chance, I'll post the "artist's summary of his approach" that will give some sense of what's meant by collaboration and the curves as speculations of possible truths.

    K?

  6. Basil says:

    I'm still a little confused even after the postings. :) Is there a statistical art contest and who can enter?