Statistics

[Eltern]

I'm trying to find a replacement for Statistica or SPSS on Linux, and after several hours of scouring the net for such a program, I've found nothing.

What I'm looking for something that has a GUI, spreadsheet capabilities, decent graphing, and a signficant amount of statistical analyses available with a few mouse clicks (namely repeated and multiple measures ANOVA). As I said, I'm basically looking for SPSS or Statistica

I've toyed around with R and R Commander, but that's still too rough of a GUI for me. It could be that I just didn't understand it sufficiently, but it seemed that I would have to do a lot of coding to accomplish what SPSS can do in three button clicks. Granted, if I were trying to do very complicated and/or unique analyses, I might like R's robustness. But that's not what I'm doing.

Is anyone else in the same position as me? What did you do? Is it possible to use a particular combination of packages with R, or some other program, to achieve this effect?

Thanks!

[akniss]

I'm trying to find a replacement for Statistica or SPSS on Linux, and after several hours of scouring the net for such a program, I've found nothing.

What I'm looking for something that has a GUI, spreadsheet capabilities, decent graphing, and a signficant amount of statistical analyses available with a few mouse clicks (namely repeated and multiple measures ANOVA). As I said, I'm basically looking for SPSS or Statistica

I've toyed around with R and R Commander, but that's still too rough of a GUI for me. It could be that I just didn't understand it sufficiently, but it seemed that I would have to do a lot of coding to accomplish what SPSS can do in three button clicks. Granted, if I were trying to do very complicated and/or unique analyses, I might like R's robustness. But that's not what I'm doing.

Is anyone else in the same position as me? What did you do? Is it possible to use a particular combination of packages with R, or some other program, to achieve this effect?

Thanks!

It seems in the free statistics software world, you either get to perform complex statistics properly, OR use a GUI... not both. R has the most functionality, but can be a real pain in the @$$ to learn to use efficiently (although its worth the time in my opinion). PSPP tries to mimic SPSS and has a GUI, but the statistical functionality is still pretty limited. There are some promising projects to put a user-friendly face on top of R, most notably JGR (pronounced Jaguar) and RKward. However, neither is in the Ubuntu repos at the current time, and I've not had success installing rkward (I haven't tried JGR).

[gbutalid]

Hey akniss, have you tried R in Dapper Drake? I'm using it in the command line, much like the RGui in Windows but I had trouble with the Graphics. I installed all the necessary x11 fonts. I also saw this problem in the R mailing list. But no one has posted a solution. I also used Octave in the command and there is no problem if I ask for a plot through Gnuplot.

The errors that I got when I try to make a histogram in R is posted below.

> hist(rnorm(100))
Error in X11() : could not find any X11 fonts
Check that the Font Path is correct.
In addition: Warning messages:
1: locale not supported by Xlib: some X ops will operate in C locale
2: X cannot set locale modifiers


Thanks in advance.

[akniss]

Hey akniss, have you tried R in Dapper Drake? I'm using it in the command line, much like the RGui in Windows but I had trouble with the Graphics. I installed all the necessary x11 fonts. I also saw this problem in the R mailing list. But no one has posted a solution. I also used Octave in the command and there is no problem if I ask for a plot through Gnuplot.

The errors that I got when I try to make a histogram in R is posted below.

> hist(rnorm(100))
Error in X11() : could not find any X11 fonts
Check that the Font Path is correct.
In addition: Warning messages:
1: locale not supported by Xlib: some X ops will operate in C locale
2: X cannot set locale modifiers


Thanks in advance.

Sorry to not be any help, but I've been using R successfully in Dapper since shortly after it was released... and I've not had any trouble with the graphics. As a possible workaround... have you tried setting the fonts to something else with a call to par()? I have no idea if it would help...

[stevetxt]

I get this error too when I try to use R in edgy. So it is not possible to do any graphics in R. There was no problem with R in dapper.

[seelenhirt]

i don't know the cause of this problem but setting LANG to something other than en_US.UTF-8 works for me.

$ LANG=en_US.ISO-8859-1 /usr/bin/R

hope this helps

m

[stevetxt]

Thanks Seelenhirt.

I've similarly been using
LANG=C R
as a temporary fix. It seems to work fine, but I still wonder where the problem comes from, so a real fix is possible. As I understand it, "LANG=C" tells R to use whatever font defaults are in the program and not try any language translation based on locale first.

[gbutalid]

Thanks!

I can now create charts. But I always start R in the commandline with this:
"$ LANG=en_US.ISO-8859-1 /usr/bin/R"

Is there a way to save this, because when I have to start R again in the terminal by typing "R" I still can't make plots--I have to type the said command again.

Thanks again for the big help.

[gbutalid]

And one more thing. When I was still using Windows and I plot something through R, I just right click on the plot and copy it as a bitmap or metafile, then paste it to a Word document. I tried it in Ubuntu, having R in the terminal, but it doesn't work. Does anyone know how I could copy a plot? Thanks in advance.

[akniss]

Does anyone know how I could copy a plot? Thanks in advance.

The best way I've found to get a plot on the linux side is to use the plotting functions in R to save them... rather than trying to copy them from the graphics device as in Windows.

png("filename.png")
 plot(x,y)
dev.off()

or similarly, using jpeg(), or postscript().