Science applications on the Mac

Ashish Ranpura wrote:
>
> In addition to experiment programming, I'm trying to find some good
> graphing packages for making publication-quality charts and graphs.
> These are typically not too complicated, but there must be something out
> there producing more elegant designs than Excel.

Personally I use PGPLOT <http://www.astro.caltech.edu/~tjp/pgplot/> for
graphics. I usually call it from perl using perl's PGPLOT module,
rarely from Fortran or C programs. It makes nice looking graphics.
Since it's been a few years since the last new version of PGPLOT, it's a
bit tricky to install on OS X. You start by telling it that it's a NeXT
box, edit out the NeXT specific parts in the makefile, and then use a
BSD specific switch at some point.

For rough and ready plots, which can be made publication quality as
well, I use gnuplot <http://www.gnuplot.info/>, which builds out of the
box on OS X.

--
Paul Schinder
schinderpobox.com

> After 2 weeks of wrestling with MATLAB and SPSS installations, I've
> decided to appeal to community wisdom.
[...]
> In addition to experiment programming, I'm trying to find some good
> graphing packages for making publication-quality charts and graphs.
> These are typically not too complicated, but there must be something
> out there producing more elegant designs than Excel.
>
> Which programs do scientists prefer for these tasks?

I think it really depends on exactly what you are doing, and what people you
are working with are using. However, I would suggest that you take a look
at R:

<http://www.r-project.org/>

It's free & open-source, and runs on Windows, *nix, and Mac (Jaguar or above
- the GUI requires Panther; no idea about Tiger). There's some limited
AppleScript support, if that's important.

If you can use MATLAB/SPSS, then you can probably work R. One of the
biggest differences is that with R (or rather the language it uses, S)
statistical analysis is done as a series of steps, with intermediate results
stored in objects. Where SPSS will give very verbose output from (eg) a
regression, R will give minimal output and store the results in an object
for subsequent work.

=Tony.Meyer

See http://tazman.princeton.edu/osx/ for a good roundup of applications for science. I run SPSS 11 under OS 10.3.9 and MATLAB under X11. Be aware that apparently SPSS will not work under Tiger (10.4) and they're planning an update (to SPSS) in the next few months.

For plotting I use Kaleidagraph, mainly because it has Windows and Mac versions and my research group has both platforms. ANother good plotting package is ProFit (www.quansoft.com).

On Wed, May 11, 2005 at 2:02 am -0700, Ashish Ranpura wrote:

>After 2 weeks of wrestling with MATLAB and SPSS installations, I've
>decided to appeal to community wisdom. Installing MATLAB under X11
>should be straightforward, but of course it's not. I haven't yet gotten
>it to work.

What sort of problems are you having? It's running fine here (X.3.8,
Matlab 6.5).

And just in case you haven't seen this:

<http://www.mathworks.com/support/tech-notes/1300/1317.html>

TimH

My colleagues all use R, which is a free version of the commercial S-Plus package. Available at http://www.r-project.org/

Hello Ashish,

For almost every aspect of my scientific work, I use Igor Pro by WaveMetrics. Using Igor, you can import and process data, and then generate beautiful graphs and images with little effort. With a little extra effort, you can control every nit-picky detail of your graphs.

<http://www.wavemetrics.com>

You can drive Igor using either the command line or the menu commands. One very nice feature is that every menu command and dialogue generates code, which you can adapt to write macros and automate data processing/presentation.

Igor also features a programming environment and several analysis toolkits, although using them does require investing a few days to experiment and explore these features.

If you do look into Igor, be sure to subscribe to the users forum. The community is very active and eager to help new users.

I hope this helps, Simon Garcia

We use Deltagraph (RedRocks software).
http://www.redrocksw.com/deltagraph/mac/

It is certainly not the end-
all in graphing, but it handles most of the standard scientific plots
that we need, and is logscales beyond Excel in user friendliness and
appropriate Mac-Centric behavior. Give it a look.
Cheers!
-David
         David D. Brand, Ph.D.
         University of Tennessee
         Department of Medicine
         http://public.xdi.org/=David.D.Brand
         http://www.utmem.edu/ctr/RDRCC_CollagenCore

Hello,

This is a bit off the topic but your "Publication quality" urged me to
mention MathMagic equation editor.
It's an Mac first equation editor that gives you the publication quality
output for the rest of us.
It worked on Tiger. There were fully functional free trials on
    www.mathmagic.com

If you need to deal with equations in your documents, please give it a try.

Hope this helps.

Bob

I run Matlab 7 (Release 14, service pack 2) on Mac OS 10.3.9. This last upgrade (SP2) was the only one that didn't cause problems. Every other installation since it re-appeared for OS X has caused huge problems for me, sometimes taking days rather than hours to solve. Almost always the problem is with the licence manager Mathworks insists on using. I've found that sometimes the best thing to do is install after restarting your Mac with the ethernet cable unplugged- this gets around some of its issues and has the advantage that you can then dare to use the software on a laptop at home or at the office. This licence manager makes legitimate users work harder than those who might have cracked copies- I'm sure the Mathworks could come up with something less intrusive and obstructive.

As for graphing, my main workhorse is still humble old Cricket Graph. It hasn't been developed since 1992, but for high quality 2D graphs and ease of use, it can't be beaten. I turn to DeltaGraph only for true 3D graphs which Cricket Graph can't make. Otherwise it is a bit of a dog. I've recently looked at Kaleidagraph but wasn't terribly impressed. Aabel (http://www.gigawiz.com/) is a program which seems to utilise the features of OS X to the fullest to produce beautiful graphs. However, to me it was mind-bogglingly difficult to get it to do even the simplest things.

Cheers,

Michael MacAskill

I can definitely recommend R for statistical purposes. It may have a learning curve, but there are several tutorials of various levels on the website. But I've sliced and diced data in R in ways I can't imagine in other programs.

For technical graphing, I'm still at a loss. I wish there was something I could recommend to PC switchers that's equivalent to SigmaPlot or Origin for easy, publication-ready scientific plotting. Not yet, although at a recent conference, both companies suggested they're evaluating Mac OS X markets.

I've gone with Aabel for much of my plotting:

<http://gigawiz.com/>

I think the user interface has some quirks and I've e-mailed the developers with suggestions, bugs, complaints, and complements. They seem like a small group, but have some great plans (particularly for linking to external programs like Matlab, SPSS, R, etc.).

I hope this market grows further. I was very happy to see a large Apple presence at the American Chemical Society meeting in March, and I'm hoping that's a sign of further interest in Mac science.

-Geoff

While I found that the early versions of Matlab on OSX were cumbersome
to install, I haven't had any problems with version 7.0.x (on mac os
X.3.5 and up, and apple's own X11). Perhaps you're trying to run an
older (6) matlab version on an older OSX version?

Should you reconsider matlab, for your experimentation, I suggest you
take a look at the Psychtoolbox (http://psychtoolbox.org/). It's an
openGL based matlab toolbox for research-grade visual stimulus creation
from within matlab. While the OSX version is still being tweaked ,
there is a large body of experience using it under OS9. There's also an
active user community and support forum.

gr.
frans



----------------------------------------------------------
Frans W.Cornelissen

I must agree with Simon. Igor Pro is great. It is also cross-platform, so that if you have to share with your windows colleagues, you can. I do as much of my data analysis as possible in this program and have written many procedures to automate standard analysis that I do. Also, my figures for all my latest publications have been done in Igor. There is a bit of a learning curve, like with any program, but once you do it, it is versatile and powerful.

Tech support by email is also very good.

Morris Benveniste

I was a beta tester for KaleidaGraph years ago, when it was Mac only.  It s the closest successor to CricketGraph, the standard at that time.  I still think it is the best, although it is beginning to look more Windows-like, now that it is cross-platform.  Earlier versions (3.6) are less powerful, but easier to use.  I've also used Prism, the preferred PC app where I work, and it has a Mac version.  Compared to Excel, both of these are in a different league, if you can spend the money.
Cheers, Gerry Waneck