Showing posts with label statistics. Show all posts
Showing posts with label statistics. Show all posts

Thursday, December 10, 2009

A Practical Guide to Geostastical Mapping

Just saw this and thought I'd pass it on:

200912101753.jpg

Tomislav Hengl has just released has new book, A Practical Guide to Geostatistical Mapping over at http://spatial-analyst.net/book/. The book is available for free browsing online and as a free PDF download, or you can order a printed copy. The book is made up of lecture material for a PHD course teaching spatial analysis using open source software. All the datasets used in the book are available at the homepage as well." Download the PDF here Interestingly, you can improve the book interactively by using an iPaper application that is part of the www.scribd.com browser. Cool idea.

From the webpage:

Geostatistical mapping can be defined as analytical production of maps by using field observations, auxiliary information and a computer program that generates predictions. The purpose of this guide is to assist you in producing quality maps by using fully-operational open source software packages: R+gstat/geoR and SAGA GIS. Materials presented in this book have been used for the five-day advanced training course "GEOSTAT: spatio-temporal data analysis with R+SAGA+Google Earth" that is periodically organized by the author and collaborators. This is an open access publication!

Tuesday, December 16, 2008

Quantitative Archaeology Wiki [Quantitative Archaeology Wiki]

Quantitative Archaeology Wiki provides a nice summary of basic statistics done in using the free platform 'R'. If you don't know R, you should - while may be not a GUI pull-down menu kind of stats package, you can literally do anything with it and do it easily (and repeatedly). And its free!

The examples shown on the wiki come from Mike Fletcher and Gary Lock's book
Digging Numbers (ISBN 978-0947816698).


Sunday, November 16, 2008

PSPP

I have updated the PSPP 0.6.1 package to include what I hope are all the libraries required to make PSPP work on the Mac OS 10.5 platform. You can download that now at: http://www.csulb.edu/~clipo/papers/pspp/

Saturday, October 11, 2008

PSPP 0.6.1

Version 0.6.1 of PSPP ( a bug fix release) was just announced at http://savannah.gnu.org/projects/pspp I created a new set of Mac OS X binaries and put them in an install package so they will end up in /Applications/PSPP (PSPP and PPSPIRE). You can download them here: http://www.csulb.edu/~clipo/papers/pspp/ Windows users should get the binaries at: http://www.gnu.org/software/pspp/get.html

For more info on PSPP, please see: http://www.gnu.org/software/pspp/

Wednesday, October 8, 2008

Stats for Students - PSPP

An anonymous comment on my previous post about "Stats For Students" suggested I take a look at the free open-source project PSPP. This program provides all of the core functions of SPSS and has a GUI that makes it much like its commercial counter part. It seems pretty functional and useful, though to use it on a Mac, one has to build it from source (and the many prereqs and pre-prereqs and pre-pre-reqs...). Windows users have it a bit easier as there are binaries for that platform: http://pspp.awardspace.com/ For those who use Mac OS X 10.5, I created a binary distribution package: http://www.csulb.edu/~clipo/papers/pspp/ for the 0.6.1 release. This is a *new* version from what was previously posted here and now includes libraries required for the software to run. The binaries and libraries are now included in the PSPP-0.6.1.zip package. Once you download it (about 86 megs due to the mass of libraries required), run the install program. PSPP and PSPPIRE will be installed the /Applications directory (in the PSPP folder). [[UPDATE]] In a previous packaging of PSPP, I somehow failed to include the actual applications. The package has been updated http://www.csulb.edu/~clipo/papers/pspp/ so that you will now get the PSPPIRE and PSPP applications installed in /Applications/PSPP (or a place of your choice). Sorry about this messup. [[UPDATE 2]] I'm uploading a new binary right now. It's uploading sloooowly as I am at the airport on wireless. But eventually there should be a functional binary available. [[UPDATE 3]] I've created an entirely new package that fixes the many problems that others had. At least it doesn't give the errors it used to. The new version is now loaded on the website. It also does not require logout (which was my mistake for not unchecking that box in the package building). <<<>>

Monday, October 6, 2008

Stats for Students

While the statistics software package available in most labs around campus is SPSS, the fact that it is not freely distributed to students (an academic version is available to students for $200 -- with your student ID you can buy it online at www.academicsuperstore.com) makes conducting independent research and working at home on stats related projects somewhat difficult. There are a number of free solutions to this problem that you can check out:


(1) For CSULB students, it is possible to freely download (while on the campus network) a copy of Minitab (http://www2.calstate.edu/csp/minitab1.asp). You can only do the download while connected to the CSULB network . This package has a similar interface to SPSS (and is produced by the same company as SPSS). It excels in descriptive stats but also has many other kinds of statistics that you will find in SPSS. It is a windows only product. I use this software myself for simple kinds of analyses.
(2) Systat offers a free version of their stats package to students (MyStat): http://www.systat.com/MystatProducts.aspx I have always liked Systat as a statistics package and use this often as an alternative to SPSS. The free version has a number of limits but generally covers all the statistics that you might use in basic research.
(3) For those who are truly geeky, I highly recommend learning how to use "R" - an open source, free statistic package modeled after S and SPlus. The range of statistics is mindblowing and you will likely find every conceivable statistic and graph that you have ever encountered (from traditional stats, to bioinformatics to spatial stats, to phylogenetics to well, just check out: http://cran.cnr.berkeley.edu/). It works on every platform (windows, mac os x, linux) and is entirely free. You can get it here: http://www.r-project.org/ It is not necessarily for the faint-of-heart when it comes to intuitive interfaces. While there are some project that offer a GUI to R (see: http://www.sciviews.org/_rgui/ I like the Java interface JGR) R works best on the command line and you have to learn its language to run statistics. However, once you do you can virtually do anything (for free!).

Thursday, April 3, 2008

Some good free Mac graphing programs

I am happily a full-on Mac users these days. I still keep a window box at home and at the office to do the kinds of things that remain only possible on a windows box due to specific software packages: ArcGIS, SolidWorks, ERDAS and well, that is about it. Increasingly I am finding new ways of doing even these things on the Mac. For example, QGIS is an excellent (and free) GIS front end to GRASS.

Some of the programs I am finding for the Mac are really gems that make using OS X a techie dream. For example (and this doesn't include the world of X Windows based software):

Plot: http://plot.micw.eu/Main/HomePage This is a great free program for doing XY charts with lots of controls for error bars, perl based scripting, etc.

Octave: http://www.gnu.org/software/octave/ An incredible package for solving linear and nonlinear problems numerically using a Matlab language - all for free. See also: http://sourceforge.net/project/showfiles.php?group_id=2888&package_id=237839

VVIDGET: http://www.vvidget.org/ The rather incredible VVIDGET application that allows one to drag and drop all kinds of features into 2D and 3D graphs. Something ya gotta see...