Skip to main content

View Diary: Statistics 101 Book recommendations (79 comments)

Comment Preferences

  •  R is powerful (0+ / 0-)

    but kind of all encompassing.  The emacs of statistical packages, if you know what I mean.   I keep trying to use it, but its not a natural fit.  I'll probably keep trying, but for most of my purposes the occasional foray into statistical analysis is well supported by the much less demanding python modules.

    # 17 days 'til the light starts to return

    by jotter on Mon Dec 04, 2006 at 11:46:47 AM PST

    [ Parent ]

    •  That's right. (0+ / 0-)

      'R' is a big powerful package with a non-trivial learning curve involved, but if you need something like it and can put in the necessary work 'R' can sure get the job done. I also use Mathematica(R) in my work, but I sure wouldn't use it to balance my checkbook when an ordinary calculator will do the job and much faster. I'm a firm believer in using the best tool for the job, which is not always the same as the most powerful one out there (although there certainly are times when such power is necessary and useful). In fact for simple tasks (like averaging, plotting, and fitting student grades) Excel's statistics functions will work just fine--although I myself actually prefer to use a cheap simple graphing package like KaleidaGraph (or DeltaGraph) for those simple statistical and graphing tasks, and reserve 'R' for the heavy-duty stuff.

      Why did we bother to defeat the Soviet Union if we were just going to become it? -- Molly Ivins

      by dewtx on Mon Dec 04, 2006 at 01:32:50 PM PST

      [ Parent ]

      •  I guess it depends mostly on (0+ / 0-)

        the size of the data sets you're working with, and how often you'll be repeating the same sort of analysis.

        Large data sets or routine/frequent analysis needs should send one over to R or another similar package, perhaps one customized for a the particular domain of interest.

        For smallish data sets I've used mostly excel in the past, but have become very interested in ggobi for quick visualizations when there are many measurements on each item.

        Another graphing package that I've found useful in an industrial setting (set up an analysis/visualization and pound it into the ground) is ploticus, which also has some built in statistics features.

        And of course, now that google is a unit converting calculator, the hand held is really an artifact of another age.

        # 17 days 'til the light starts to return

        by jotter on Mon Dec 04, 2006 at 01:59:49 PM PST

        [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site