Page 2 of 13 FirstFirst 123412 ... LastLast
Results 11 to 20 of 122

Thread: Scientific Software options for Ubuntu

  1. #11
    Join Date
    Nov 2004
    Beans
    59
    Distro
    Ubuntu Development Release

    Re: Scientific Software options for Ubuntu

    Re: Akniss and the R-project.

    I see the power of the R-project, but the interface is horrible. Is there anyone working on a good interface?

  2. #12
    Join Date
    Oct 2005
    Location
    Wyoming, USA
    Beans
    484
    Distro
    Ubuntu 12.04 Precise Pangolin

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by ericsp
    Re: Akniss and the R-project.

    I see the power of the R-project, but the interface is horrible. Is there anyone working on a good interface?

    I suppose that really depends on what you are used to. In my field, SAS is the standard. This makes the switch to R rather easy, as they both have a command-line programming language feel to them. Most people who feel as you do typically come from Minitab or SPSS or... Excel *shudder*. The learning curve is steep; however coming from SAS I've really enjoyed the transtition. I can make a plot in R with two or three lines of code that would take me a page and a half of code in SAS. However, two or three lines of code is two or three too many for those used to pointing and clicking their way around Minitab or SigmaPlot.

    The true "power of the R-project" really comes from the fact that it is a full progamming language. We are not limited to a set of default analysis methods which may or may not be appropriate for the data. I have written a few functions in R that allow me to create detailed diagnostic plots with a one-line command. This sort of thing is just not possible with 95% of commercial gui-based statistical packages. It can be cumbersome at times when one only wants to perform simple analyses or transformations on a data set; but if this is all you want out of a statistical package, then R is probably not for you anyway. Any attempt at simplifying the interface to point and click will necessarily take away functionality, as there is no way to incorporate all the functionality of R into a few context menus.

    With all that being said, there are a few projects out there that have put a gui interface on top of R. Rcmdr is available in the Ubuntu repos and does an ok job. The one that looks the most exciting in my opinion is JGR (pronounced Jaguar). It is java based and so platform independent. I have not personally tried it but have heard good things.
    This looks nothing like my signature...
    My Page

  3. #13
    Join Date
    Oct 2005
    Beans
    21

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by ericsp
    Re: Akniss and the R-project.

    I see the power of the R-project, but the interface is horrible. Is there anyone working on a good interface?
    This is by far the most promising thing I've seen. http://kde-apps.org/content/show.php?content=14880

    You can also install a GUI called Rcmdr within R. Just connect to an R reposititory, install the package, and at the input prompt type

    Code:
    library(Rcmdr)
    It's not the prettiest thing in the world, but it's stable.

    One other thing. I've recently begun experimenting with using DOSbox to run an old DOS version of Kwikstat. It works well after some fiddling with the install (edit a script so it Kwikstat doesn't look for a printer on install). This might be an option for those of you who are still using older software. I find Kwikstat quite satisfactory for most things.

    Aside from statistics, I use CXoffice to run Endnote. I just cant switch. There is a Linux version of Maple, so no problem there. Maxima is also a decent free computer algebra system.

    I really wish someone could get a recent versionn of SPSS working in Linux with a GUI.

  4. #14
    Join Date
    Oct 2005
    Beans
    1,887

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by ericsp
    I see the power of the R-project, but the interface is horrible.
    check out this: https://stat.ethz.ch/pipermail/r-dev...ch/036728.html

    This is a bug (or, rather, a feature request) I filed a while ago. You will find your response in devels' replies to my bug report / feature request......

  5. #15
    Join Date
    Jun 2005
    Location
    Melbourne, Victoria, Australia
    Beans
    4
    Distro
    Ubuntu Breezy 5.10

    Re: Scientific Software options for Ubuntu

    I'm forced to use Stata at work but I'd like to learn R. Do you think going from Stata to R would require a huge mental leap?

  6. #16
    Join Date
    Oct 2005
    Location
    Wyoming, USA
    Beans
    484
    Distro
    Ubuntu 12.04 Precise Pangolin

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by claudine
    I'm forced to use Stata at work but I'd like to learn R. Do you think going from Stata to R would require a huge mental leap?
    I don't know much about Stata... From looking at their website, it looks like it is predominantly a command based package as well. (I'm actually kind of interested in playing around with Stata now!) I think the switch to R is much easier for those of us who are already used to typing in commands to get means, regressions, and box plots.

    However, learning any new language is frustrating at times. It took me a couple years to learn SAS to the point that I was reasonably proficient, and it was difficult to transition from the proc ****; world over to the function() world. I gave up a few times, but eventually got the hang of it and am now much more efficient in R.
    This looks nothing like my signature...
    My Page

  7. #17
    Join Date
    Oct 2005
    Location
    Wyoming, USA
    Beans
    484
    Distro
    Ubuntu 12.04 Precise Pangolin

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by Kyle-
    This is by far the most promising thing I've seen. http://kde-apps.org/content/show.php?content=14880
    Kyle, that looks fantastic! I hadn't heard of it yet. I suppose the chances of it getting into the Dapper repos are null...
    This looks nothing like my signature...
    My Page

  8. #18
    Join Date
    Oct 2005
    Beans
    1,887

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by akniss
    Kyle, that looks fantastic! I hadn't heard of it yet. I suppose the chances of it getting into the Dapper repos are null...
    try suggesting it to the backports (there is a forum section for that). Let us know of the thread link so we can support.

  9. #19
    Join Date
    Jun 2005
    Location
    Melbourne, Victoria, Australia
    Beans
    4
    Distro
    Ubuntu Breezy 5.10

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by akniss
    I don't know much about Stata... From looking at their website, it looks like it is predominantly a command based package as well. (I'm actually kind of interested in playing around with Stata now!) I think the switch to R is much easier for those of us who are already used to typing in commands to get means, regressions, and box plots.

    However, learning any new language is frustrating at times. It took me a couple years to learn SAS to the point that I was reasonably proficient, and it was difficult to transition from the proc ****; world over to the function() world. I gave up a few times, but eventually got the hang of it and am now much more efficient in R.
    Yes, Stata is very much command-driven. I'm writing Stata scripts in Emacs with ESS mode. Stata mostly does what I need, but it's not free (my philosophical objection). And all the statisticians at work use it and seem to think R is too "difficult". It seems the main issues with R are its object-oriented nature and syntax.

  10. #20
    Join Date
    Oct 2005
    Location
    Wyoming, USA
    Beans
    484
    Distro
    Ubuntu 12.04 Precise Pangolin

    Re: Scientific Software options for Ubuntu

    Quote Originally Posted by claudine
    Yes, Stata is very much command-driven. I'm writing Stata scripts in Emacs with ESS mode. Stata mostly does what I need, but it's not free (my philosophical objection). And all the statisticians at work use it and seem to think R is too "difficult". It seems the main issues with R are its object-oriented nature and syntax.

    R certainly does have a cryptic syntax... however, I would argue that there is nothing more cryptic than Emacs , so if you've mastered that... ESS is also able to interface with R, so I think it wouldn't be too terrible a transition for you. I gave ESS a whirl a few months ago and just coulnd't get hooked. I thought it was ok, but decided KDE's kate editor was better suited for my needs. I guess there's only one way to find out: install it and give it a try. Just take it slow and don't think you'll be able to master it the first few times. If you're anything like me, it will take a while to 'unlearn' your other language (It took me about a month to stop putting a semi-colon at the end of each line). Good luck, and if you have trouble getting started there are pretty good books, online help, and documentation available. And I'm sure there are a few R junkies on this forum besides myself as well!
    This looks nothing like my signature...
    My Page

Page 2 of 13 FirstFirst 123412 ... LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •