Latest Release: 23 October 2003
Rapid statistical analysis of scientific data under the GPL
Other statistics packages for Linux:
R (free version of S-Plus)
Latest Development News
22 October 2003
New release with some minor bugfixes!
11 September 2003
New release made today (codenamed "Texas!"). With this version you can transformation your data! There is also the basis for the native sml format, but it is not operational yet. It can open some very limited files, but the save file routine has yet to be worked out properly. It's a longer process than I anticipated, but I hope to have it operational very soon. Have fun!
02 September 2003
Oops - I've just realised that I haven't updated this page in a *long* time.
Latest news: I've been working on the "transform data" stuff which seems to be working alright! The interface isn't as good as I hoped, but it works. The user can enter their own functions (using Python syntax), or they can use one of the standard functions provided just by pressing on a button. The transformed data are saved into a new column which is given a meaningful name.
The package has been looked over, and a lot of changes have been made "under the hood". The code has been cleaned up a little, though no new tests have been written. I am still working on the multi-factorial anova code, but should (hopefully) be getting some help from a statistician very soon.
I am also working on an XML data format: this will record all the data, all changes to the data, all analyses and all the output. This will enable a user to completely audit their analysis session from the very start to the very finish, and everything will be included into one complete file.
None of the above has been released yet, but I am anticipating making a new release very soon!
05 November 2002
New release made. This incorporates a self-installing file for Windows, an all-inclusive download for Linux (no dependencies - I think!), scripting support (to write your own tests), extremely basic bar charts of means, interface updates, printing of output, some more tests, improved help pages, new users manual, and lots more.
Scripting support! Although a release has not been made with this feature, it is being developed and tested, and will be included in the next release. The scripting support allows access to the statistics functions in SalStat using the Python language, and will allow users to develop their own tests using these functions, and of course, the rest of Python is available too.
03 August 2002Fixed a bug in the between subjects anova and added the Cochranes Q non parametric test
02 July 2002
New release of SalStat - the code has been re-written in an object-orientated manner. The code base is now much smaller and faster.
11 June 2002
New release of SalStat containing help pages - the user manual has been converted into html pages which are displayed using the help window.
06 June 2002
New release of SalStat. The Friedmans test is coded and briefly tested (please test against existing packages if you can and report back!), some online help documentation is included, and the save/load data routines have been improved. Also added the 2 sample sign test and the f-test for variance ratios.
01 June 2002
New release of SalStat today which is a consolidation release. Not much has been added outside of a few bug fixes, and the addition of a couple of brief help pages.
05 May 2002
I have been informed that SalStat can run on Mac OSX! Any prospective users will need to follow the instructions in the README file submitted by Ludger Humbert. Also added are the single factor analysis of variance (within subjects) with results tested against SPSS, and the 1 sample sign test (non-parametric 1 sample t test). SalStat can also load comma separated value datafiles (where the data are separated by commas, and not just spaces or tabs).
29 April 2002
Some of the tests have been tested - the t tests work! Between subjects single factor has been added, and more improvements to the GUI have also been made.
19 April 2002
There was a small bug in the Spearmans test which I have fixed (still needs beta testing folks! Hint, Hint!), so it should be more reliable. There is now a dialog for 3+ sample tests (single anova, Kruskall-Wallis), but it's not hooked up to the algorithms yet. The output is more nicely formatted (to 3 decimal places - I must implement a way for the user to change it).
The html output can now be saved which saves a bit of work - it still doesn't copy to the clipboard yet as the components I have don't allow it - although I could easily implement a way tocopy the whole lot to the clipboard which I suppose is better than nothing.
The median and mode have been added to the descriptive tests available, and the toolbar is a lot better than it was (and is actually functional). Also, when a data set is loaded and is larger than the grid, the grid automatically resizes itself to be slightly larger than the dataset (4 or 5 rows or columns each way).
16 April 2002
It now does some tests! There are t tests (related & unrelated samples), correlations (Pearsons, Spearmans, Kendalls), Mann-Whitney U, Kolmogorov-Smirnov, Rank Sums and Wilcoxon t. Don't trust them yet - they haven't been properly tested! YOU HAVE BEEN WARNED!!!
Also, I have put up 2 archive, one the traditional tar.gz, the other a zip file which is probably more portable.
13 April 2002
SalStat can output a range of parametric descriptive statistics so it is partly functional. I have also added the correlations options (non-functional just yet!), and the html output works fine! I'm also thinking of using Gary Strangmans stats.py module - it contains lots of tests, and all I have to do is modify the output so that it does html (and not command-line), incorporate it into the interface, and away we go!
11 April 2002
The functionality of the data grid has been improved: Columns and rows can be deleted or added. Data can be saved and loaded (in ASCII text format), but this needs to be further tested to ensure it works properly. Variable names can be entered and used, and there are dialogs for most of the tests. No statistical analysis is possble yet, but the project has only been going for a few weeks - hey, what did you expect??? Give it another month or so, and you shall have t-tests!
02 April 2002
Started the "proper" web site with downloads, screenshots and so on! Development is concentrating upon the interface - once the framework is complete, adding the "working" code is relatively straightforward (I hope!)
The dialogs for one and two condition tests work, the About dialog sort of works, and a new output sheet can be called easily. The grid has some manipulation (adding/deleting rows & columns). I have also worked out the code to size the data set. This is useful for the analysis and the saving of data.
To Do: Adding more grid functionality (adding n rows and columns), saving and loading data as ASCII files, functionality for the variables grid (variable names, missing data etc), and a routine to calculate the used columns for the dialog boxes.
This project is hosted by SourceForge.net.