Tag: R

R , Ubuntu, RCmdr Updates

Some nice updates for R followers-

1) Rather than have an Icon for R – There is a seperate icon for RCmdr in Ubuntu Karmic Koala – Thus the default screen on opening is R Cmdr.

2) REvolution Computing has managed a coup with their bundling of their libraries with the R Distribution in Ubuntu Karmic Koala( see screenshot). We however still are waiting for who gets the credit for that ( Daneese Cooper or the long suffering Mr Smith)

Screenshot

 

3)Karmic Koala offers 2 GB free space  for storing data in the cloud for every user and 50 GB at 10$ a month. This helps with your storage costs. Data is protected thanks to an oauth login id and machine specific tie-in.

4) RCmdr has a great new plugin for DOE (Design off Experiments) students. DOE is a powerful and under utilized technique especially in Web Analytics. This is promising given that Dr John Fox ( whom we interviewed on this website) has going on ahead and seems clearly to have established RCmdr as the introductory GUI for beginners to R.

(see screenshot 2 below)

5) The Karmic Koala is very easy to install and very intuitive to use- Don’t want to give up your Windows ( well just install a dual boot which takes less than 1 hour on a fast internet connection or 15 minutes if you have a DVD)

 

Screenshot-1

6) What are other Statistics softwares doing? If they are not too keen on helping Microsoft get more sales ( especially student OS licenses) why don’t they offer the Ubuntu version free for students ( and besides once and for all put to rest the open source credential controversy)

 

Decisionstats Interview at Big Data Summit, AsterData

 

For a change, I got interviewed at Big Data Summit, sponsored by Aster Data. With a special thanks to Tasso , Steve Wooledge and his team, and the lovely Michelle from http://www.zagcommunications.com

Note this is just the raw unedited footage so it can be pretty repetitive, so if you want the final interview it should take some time. Anyways I managed to not make a complete mess of it. Have a look.

Running Stats Softwares on Clouds

If you have a small beatup laptop and want to rent say heavy hardware and expensive software, a technology demonstration is underway at www.analysis.utk.edu

We basically use a Citrix Server to run R, SAS and JMP ( among others on the Web through the browser)

Note- If you just run this on a normal Web Server with lots of hardware packed behind it- you can start giving cloud computing solutions to your clients for free.

And eliminate the OS ( Windows unless you like them ) and HW ( Read HP etc etc)

thus reducing Total Cost of Ownership for your final customers.

Is Revolution / SAS / SPSS Listening.

See Screenshot-

(Note this can be of terrific use for say software companies wanting to license out new markets in Asia as they can also analyze usage data and share the efficiencies with newer users. By adopting a Software as a Service Model ( by optimizing the revenue stream for cannibalization effects) they can also gain an advantage over more established players ( like R and SPSS /IBM can do with its need for more hardware and broader sales and distribution network)

Note- these are my personal views only and dont represent the University’s views. For more on University of Tennessee ‘s technology initiatives please go to http://oit.utk.edu/index.php where Scott Studham is using his expertise to revolutionalize the way education costs can be lowered using technology.

sas

Big Talking

Here is a link to my video capture of classroom presentation I did for some students today- ( note plays on Windows Media Player only and uses streaming thanks to the University of Tennessee’s high tech initiatives )

The slides are given in the form of SlideShare presentations and also here-

http://docs.google.com/present/view?id=dcvss358_955fsvm8mfr&autoStart=true&loop=true

and

http://docs.google.com/present/view?id=dcvss358_961djjqdtdd&autoStart=true&loop=true

The topics are

1) Creating an R Package easily ( till first 30 minutes)

and

2) Offshoring 101 (last 25 minutes)

Title: R and Offshorring
Speaker: Ajay Ohri
Desription: Oct 22, 2009
URL: http://vcweb.bus.utk.edu/20091022-124732-cap530/

ps- I am sorry for the accented language and hope audio is audible.

Another R Tutorial

This one is a work in progress but it can be used as a generic tutorial for creating and publishing R packages and then running them in an HPC environment. It can also be used for translating existing Algorithms in R or creating new Statistical Algorithms in R.

http://docs.google.com/present/view?id=dcvss358_955fsvm8mfr&interval=5&autoStart=true

High Performance Computing within R: Tutorial

The best tutorial for fast and easy learning of parallel computation, code optimization and general High Performance Computing is by Dr. Dirk which he gave at User9 . He is a nice guy with a pleasant manner and temper though we disagree on his choice of Debian and my choice of Ubuntu as Linux.

Citation-
http://dirk.eddelbuettel.com/papers/useR2009hpcTutorial.pdf and download it useR2009hpcTutorial

You can also read his blog here

http://dirk.eddelbuettel.com/blog/

His writing style is well, he writes interesting Linux is all I can say.