55

I am currently doing some experimental work and I have a lot of data to trawl though. I use Gnumeric, and it's very good, but often I feel there has to be something better.

Ideally I would like the maximum number of features with a minimal learning curve, but really I'd just like to know if there is something better than Gnumeric that I can use for manipulating and plotting data.

What would you recommend?

Eric Carvalho
  • 54,385
Helix
  • 2,363

15 Answers15

33

I am a Physics student and found that the best scientific plotting software for Ubuntu is QtiPlot. It is very similar to Origin, and works really well.

  • 3
    QtiPlot is open-source, you can download the source from here: http://prdownload.berlios.de/qtiplot/qtiplot-0.9.8.3.tar.bz2 Like all open-source programs, you are free to edit this source code and compile it. However, the author does not offer free, up-to-date, compiled binaries. For this, you have to subscribe to a maintenance contract. I don't know if this is a good practice, but the program is worth the cost, and if you can't or don't want to afford it, you are always free to compile by yourself or to use the binaries that comes with any Ubuntu. – nicocarbone Jan 06 '11 at 00:18
  • 1
    it is fully floss.fsf fully supports selling free software. – Lincity Feb 20 '11 at 10:30
  • however, qtiplot gets very very slow on ubuntu if the size of the data in a matrix is large, say 1K. Google says its a bug registered on launchpad. Any solution to this problem? (none given on launchpad) – Pushpak Dagade Apr 08 '12 at 11:03
  • There only seems to be support if one has a maintenance contract. I asked the author if it is possible to remotely control QtiPlot like it is possible with Veusz and he asked back if I have a maintenence contract. Without contract no answer... He also told me that there is no official community forum. I also pointed him to the many unanswered stackoverflow questions about QtiPlot and he answered that he "has nothing to do with that forum". On one hand I can understand the business model. On the other hand, I thought a "simple question" regarding the available features would be answered. – Stefan Dec 18 '15 at 16:40
  • @Stefan and nicocarbone do you know if it's possible to open Origin projects with qtiplot? The qtiplot available in the Software-Center does not support opening origin projects, sadly. I'm wondering now if I should pay the 20eur for a single user license. – iamatrain Dec 21 '15 at 10:56
  • I did not try if it works. However, at least in the opening window (of the windows version) there is an entry for the origin project file extension. – Stefan Dec 22 '15 at 16:42
27

Numpy and Matplotlib make a good combination for processing and displaying data.

Rudy Matela
  • 103
  • 4
  • +1 I used Matplotlib while making scientific report in the nuclear industry. Bonus : it is free software and Python offers you infinite possibilities.
    I do not say Matplotlib is trivial to master but it is really worth learning it. You have a good official documentation and could also get good support on stackoverflow.
    – Maxime R. Jan 04 '11 at 20:42
21

I would suggest Gnuplot. It has a nifty set of features and is good documented. So if you take some minutes to skim through the documentation you'll get the basic idea. I use gnuplot for nearly all my plots, only when I don't need the full set of features I tend to use TikZ from LaTeX.

qbi
  • 19,125
19

The R language is also quite popular and can be combined with Sweave for use with Latex.

11

Ggplot2

Is one of the best data visualization programs available. It implements the ideas of Edward Tufte, author of classics in graphical design and scientific communication as 'Beautiful Evidence' and 'The Visual Display of Quantitative Information'.

The Deducer GUI makes it possible to use ggplot2 without requiring knowledge of the R programming language in which ggplot2 is implemented. If you can use excel, you can use Deducer. Your statistical analyses will be valid and your graphs (thanks to ggplot2) will be effective and pretty.

#dependencies
sudo apt-get install r-core
sudo apt-get install rJava default-jdk
sudo R CMD javareconf
sudo R
#to install deducer
install.packages('JGR')
install.packages('Deducer')
library(JGR)
JGR()
#in JGR
library(Deducer)
8

Sage might be good for that. It ties together a lot of open source math tools to create a very extensive and flexible app.

DLH
  • 2,445
  • Sage is nice because it is open source and uses Python as a scripting language. The setup is similar to Wolfram Mathematica, which is also available in a linux version but the cost is fairly high. – GaRyu Feb 01 '13 at 12:07
8

I've used qtoctave. It is similar to MATLAB if you've used that before.

You can install it from the repositories: sudo apt-get install qtoctave

jumpnett
  • 5,975
5

gnuplot and xmgr/grace are probably the oldest Unix scientific graphing programs. I still use gnuplot from time to time (BTW it is not GNU and some consider it not free), because I know it and I've been using it for many years, but it hasn't changed much in this century and it's not user-friendly according to today's standards.

I think that the most promising programs now are QtiPlot, LabPlot and Veusz. The first two are similar to Origin (the most popular plotting software on Windows). QtiPlot has a full-time developer and it seems to be developed more actively. Veusz is different than Origin clones and unlike other programs it is written in Python. It's not in the distro yet, but it has PPA.

Another program that I use for plotting data is fityk. It is specialized in curve fitting and I use it for plotting mostly because I know it well (I wrote it), but I guess in most of cases QtiPlot or Veusz will be the best choice.

marcin
  • 151
4

I used SciDavis, Scilab and MatplotLib. However lately I am using ParaView, but this is not an easy to use program. The previous ones are easy.

Misery
  • 3,484
2

MagicPlot is also available for Linux, it requires Java. It is very useful for creating good-looking graphs and some processing. And it is free for students.

Alexander
  • 121
2

I would suggest DataScene. It produces really cool graphs and chart animations. I found the learning curve is fairy flat because of the Wizard and tutorials. You may find more information on DataScene at:

http://www.cyber-wit.com

Jeff
  • 21
2

Veusz is the best open source plotting tool I could find so far. It allows to set very detailed attributes of scientific plots, like minor and major tick size. It also provides operations to manipulate data sets. It supports SVG export and can be remotely controlled from other programs. Furthermore, my experiences with the support have been very good. The author answered my question within a day and implemented a feature request within two weeks.

muru
  • 197,895
  • 55
  • 485
  • 740
Stefan
  • 119
  • 3
1

MATLAB might be the best but it's not only for plotting and it's not free (actually it's expensive however if you are student, you can probably get it from your school).

0

R would be best for both statistical tests and graphs. If you're fine with programming, go for R. It's open-source and powerful.

Or give BioVinci a try if programming costs you too much time. It lets you drag and drop your data to run statistics and create plots. I like the modern plot types it offers, like violin plot and interactive 3D scatter plot (with hovering info). Plus there's PCA -- really helpful for scientific research. One more, it supports Ubuntu 16.04, 18.04, and Debian 9.

Hope this helps! Here's a screenshot of its PCA 3D plot.

-1

I would like to suggest the supermongo for scientific use. Although it is expansive but you can get it from your institute or research center. This is very user friendly and easy to operate. You can plot your data with high resolution and advanced settings.