Re: How good is R?

In summary, my advice is for you to work with R.
The important thing is to learn about statistics.
If you have to use some other packages later, no big deal.

casioculture@xxxxxxxxx wrote:

> How does R compare to SPSS?

SPSS was invented back in the bad old days, when programming
was hard and, by gum, they didn't want you to forget that.
So the main goal of the SPSS user interface is to keep people
away from the command language (a clumsy mix of Fortran and
Cobol) by providing menus and dialogs to carry out every task.

I don't like GUI's so that doesn't appeal to me.
On the other hand, R is a joy to work with at the
command line. R is widely used in statistical research,
because it is so flexible and powerful.

SPSS and SAS have targeted analysis of large-scale data
sets. These are probably mostly commercial data sources
(customer databases, transaction records) with 1 M to
10 M or more records. SPSS and SAS have also created tools
within their own systems so that you don't have to cobble
together an end-to-end solution out of different software
to do each task. If you're doing large scale commercial
analysis, and the task is well-specified from the outset
(i.e., no research) then SPSS and SAS are worth consideration.

That said, I've worked with R (as part of a larger
system) on data sets of 100 k to 1 M records and it
works fine.

Steer away from spreadsheets of any type.

For what it's worth,
Robert Dodier

.