Drawing Graphs

[takakia]

Can Gnumeric make barcharts showing error bars? I tried OOcalc, but it couldn't so I had to do it in excel and use the graphic in LaTeX.

[xadder]

Don't know about gnumeric, but matplotlib has this. Check out the barchart_demo.py in the examples directory.

[xadder]

In reply to befana, I'd also recommend matplotlib - it has the same feature of showing the x,y values of the mouse point.

As may be obvious by now, I like matplotlib!

[tweedledee]

Can Gnumeric make barcharts showing error bars? I tried OOcalc, but it couldn't so I had to do it in excel and use the graphic in LaTeX.

Gnumeric has decent handling of error bars, much better than OO, and I think somewhat better than Excel, though not as strong as dedicated plotting programs (e.g., matplotlib). Gnumeric also hs the advantage of very little in the way of learning curve if you're used to spreadsheets and are in a hurry.

[befana]

@ xadder Thak you for your reply. You're happy to know how to work with matplotlib!
I need to make many graphs with data from exel (it is because I want to completely remove my Windows installation, but have to make these graphs for my PhD thesis), and this data is really big amount - each graph has about 5000 rows, and fortunately, only two coloumns. Is it possible to make this with matplotlib?

[xadder]

@befana,
matplotlib (MPL) will easily handle data arrays of 5000x2. I don't know its limits, but I haven't hit them yet and I work with much bigger arrays.

I don't know how to get from excel into MPL though, as I have never needed to import from excel. This would be a job for the python script I think, and I am sure it has been solved years ago. Alternatively, exporting from excel to plain ascii (space or comma separated) should give a file which is easily loaded into python/MPL. If in doubt, a query to the MPL support list will likely give answers quickly.

I didn't know python at all until I wanted to test MPL, but I found it rather easy to learn both python (noon-style) and MPL to get things done. MPL needs a little effort, but if you want a powerful plotting framework (something like a MATLAB replacement than an excel one) then I think it is a very good choice.

[[woodstock]]

Hi xadder!

could you please post some links to tutorials that you found especially helpful for beginners (i.e. folks that know neither python nor mathplotlib).

Thank you very much.

[xadder]

Hi woodstock,

Don't need a link. Type matplotlib into google, and go to the cookbook/wiki entry. A bar plot with error bars is the 3rd example of the simple plots. (And actually, the code they give looks more complex than that in the examples directory I got with my matplotlib installation - there are many styles of working in MPL though).

The tutorial and user's guide are very good also to get the basics, although for advanced stuff and details then working through the examples directory (installed with MPL) seems to be the way to learn.

MPL does need some work, but looks like a good long-term investment. Same as python or perl

I installed all of this through aptitude or synaptic on different machines. No problems that I remember.

(And I am not saying that MPL replaces gnumeric or Rlplot - all have their place and I wish I knew them all well.)

[takakia]

Gnumeric has decent handling of error bars, much better than OO, and I think somewhat better than Excel, though not as strong as dedicated plotting programs (e.g., matplotlib). Gnumeric also hs the advantage of very little in the way of learning curve if you're used to spreadsheets and are in a hurry.

Thank you tweedledee, Gnumeric had really what I wanted(able to add standard deviation to bar charts). It was quite fast and easy too. But I had to save graphic as SVG, convert the SVG to .esp using Inkscape and includegraphics in LaTeX. Worked Well. .So I had no need to try Matplot.

[nour02]

Ploticus is very good too. It's on command line.