As far as the typing, data analysis, graphing, etc
It will depend very much on what the journal they hope to publish in uses.
Lots of academics will use TeX, or variations such as LaTeX.
You should hunt around on the net and find the version and variation that most closely satisfies your needs. The big reason academics gravitate towards it is that it is "free." They view their grad student's time as being an infinitely malleable cost-free resource. So if it takes a few hours, or days or weeks, to get LaTeX working just perfectly, they don't mind at all. Grad students do not have lives outside the lab.
Other things people use will be fairly mundane. In our office we are pretty much MSOffice bound. This is because our clients are mostly MSOffice bound. I use MS Word but hate it. I use MS Excel and find it has some good things and some things that chap my behind. I use Powerpoint, but find it usually sucks every drop of moisture out of a presentation.
Lots of people use Matlab and various other tools of that nature.
For more serious work, there will be custom tools.
You publish exported Excel graphs?
In place of "publish" insert "send to the client." In my line of work there isn't much publishing going on.
But, yes. Copy-paste into an MSWord file, and off to the client it goes. There are some places you have to do some hand-nudging. Some things mess up fairly predictably when you do this. But that's what we do.
Well, he didn't ask about that.
Actually, he didn't ask about anything. He didn't actually ask any questions at all. He was too lazy to put the question in the body of his post.
However, I routinely send documents to the national regulator. I'm going to count that as a "publication."
Data analysis: ROOT C/C++ custom scripts. Fortran model code. A bit of perl, a bit of python, a bit of Mathematica.
Graphing: For a publication, ROOT and gnuplot. But sometimes also matplotlib (python) and a bit of Mathematica.
ETA: This is me personally. Other people will use totally different things. You use the right tool for the job.
For many years we used Gnuplot exclusively, but we needed something simpler for younger, less experienced co-authors, so we've gravitated toward graph.exe.
We do a lot of analysis in the OpenOffice spreadsheet, Calc, I think. Some of the graphs are done in there, a few co-authors have learned to customize and make things really nice.
There's a program called Scidavis that is easier to use than Gnuplot, but also gives the parameter uncertainties for least squares fitting.
We roll our own Fourier analysis code.
SPSS, MATLAB, R, but you'll see a TON of GraphPad Prism. It is very easy to use and is pretty much written by scientists for scientists.
Separate names with a comma.