1. 14
  1.  

  2. 2

    Currently, I’m using matplotlib quite extensively for my current research. This article is definitely interesting and I know that matplotlib also it quite the beast from looking at many examples and generating figures and animations over the past month.

    I find it interesting to see how the data is being brought to the screen and how much bias is actually going in there by simple finetuning how it looks. I know that I am guilty of it and I had to ask myself whether what I wanted to chaieve was really necessary.

    Anyways, thanks for that overview, although I hardly use R, visualization libraries often serve an intersting purpose because they have to accomodate so many use cases.

    1. 2

      I find it interesting to see how the data is being brought to the screen and how much bias is actually going in there by simple finetuning how it looks.

      Oh yes, definitely. There are a lot of choices in visualization and practitioners always are trying to choose the best angle, maybe sometimes even overstating their results. I feel this is a broader problem that touches more than just visualization.

      Anyways, thanks for that overview, although I hardly use R, visualization libraries often serve an intersting purpose because they have to accomodate so many use cases.

      I know there are people who only use R for visualization. Maybe that’s why this particular article reached >10 upvotes. A few other R articles I posed were at ~3.

      One thing that (in my opinion) keeps R from being used more is its inability to return images straight to stdout. You always have to do the

      png(filename, width, height)
      plot(...)
      dev.off()
      

      rigmarole. Which is not so convenient. I wish things like these:

      command ... | R -e "plot()" > file.png
      

      were more polished on the R’s side. In other words I think R has (had?) a potential to be the perl of graphics. But it lacks a convenient stdin/stdout.

      I have no experience with matplotlib - how does it handle these things? I know gnuplot is convenient from the command line, but never looked into it too much so not sure how powerful it is.