Why LinkedIn and Twitter are up for grabs in 2012-14?

Given Facebook’s valuation at $60-$100 billion , Apple’s $100 billion cash pile, Microsoft’s cash of $ 52 billion, Google’s cash of 43 billion $ , there is a lot of money floating. I am not counting Amazon as it deals with its own Fire issues.

But what is left to buy. In terms of richness of data available for data mining for better advertising- it is Twitter and LinkedIn that have the best sources of data.

and LinkedIn is worth only 9 billion dollars and Twitter is only $8.5 billion dollars. Throw in a competitive dynamic  premium, and you can get 50 % of both these companies at 13 billion dollars. if owners dont want to sell 100%, well buy a big big stake.

Makes a good case- buy the company- buy the data- sell them ads- sell them better products.

What do you think?

Third Party Extensions to WPS

WPS, which uses the language of SAS, and is much more moderately priced- recently announced Version 3

As per Minequest-http://www.minequest.com/

Prices start from $1206.

WPS  is a SAS language alternative application that can run most of your existing Base
programs without modification. WPS also provides support for a limited subset of
graphics via WPS Graphing and statistical procedures via WPS Statistics.

In addition to the broad language support, WPS is affordably priced and much less
restrictive in its licensing. WPS supports many elements of the language including
the macro language facility, the data step, most of the PROCS found in base, and
includes access engines to ODBC, DB/2, Oracle, MySQL, OLE DB, Informix,
GreenPlum, Sybase, Teradata, SPSS and other database systems.

WPS has a feature that allows you to develop extensions for it.

http://teamwpc.co.uk/products/wps/modules/language_sdk

Develop Custom Language Items

Anyone who is familiar with traditional programming languages such as Assembler, C or C++ can use the WPS Language SDK developer module to add language items to extend the language of SAS support in WPS.

Once you have created your own custom language items, you can freely redistribute them to anybody who uses WPS on the same platform.

Some existing third Party extensions are-

http://www.minequest.com/Bridge2R.html

Bridge to R for WPS
The Bridge to R is an application created and developed by MineQuest that allows the WPS
programmer to access the R system. With the Bridge to R, WPS Users can write and execute R code
from within the WPS Workbench or in batch mode using WPS. A simple yet powerful integration facility
allows the WPS developer to write R statements to access advanced features found in R, test
algorithms and create new statistical methodologies.

One of the greatest advantages of the Bridge to R is that you and your developers can use the SAS
language for manipulating and reporting your data and use R for advanced graphics and statistical
computing. The Bridge to R allows WPS Users to write R code and execute R code from within the
WPS Workbench and receive the results back in the appropriate log and listing window. The Bridge to
R even supports running your WPS and R programs in batch for production purposes.

It allows you to use either 32-bit or 64-bit R depending on whether your OS platform supports 64-bit computing. On 64-bit  platforms with sufficient RAM, developers can process large amounts of data using R.

The Bridge to R requires WPS 2.5 or higher. The Bridge is provided as part of the Analyst Power Package
and is included when you license WPS from MineQuest.

http://www.minequest.com/WindowsPowerPack.html

Parallel Process with MPExec
MPExec allows the WPS developer to dramatically reduce processing time by easily implementing parallel processing on their Windows workstation or  server. MPExec allows WPS developers to take existing WPS/SAS code and by inserting a few statements, allow your programs to access all the cores  and resources on your Windows hardware platform. MPExec allows you to parallel process up to 255 tasks and manages the collection of log and listing  files for you.

Other 3rd party applications are-

http://www.uniqcus.com/english/products.html

UniQZConnect

UniQZConnect will allow your SAS and WPS users to have a SAS or WPS session on a Windows Workstation to connect directly to SAS or WPS on z/OS. This give the user the ability to download and upload data from and to z/OS, but also to submit SAS jobs for remote execution.

The product support Wizard user dialog, full script interface, and a mixture of theese.

The Performance of the product is only limited by your current mainframe bottle necks and the current transfer limits of your FTP connection. The Security is handled through FTP ensuring that you are in compliance with your current security policies.

You can download a product sheet for UniQZConnect here.

and an upcoming application from Wolfgang at http://www.wcmat.com/cmat/

Things ToDo
CMAT was first written for Unix and later for Windows. We are now working on a 32 bit version for Linux and Unix, and after that on a 64 bit version for Windows, Linux, and Unix. We are also working on an interface to WPS and R.

—–

Overall creating applications or extensions can help WPS reach a wider audience. While Rapid Miner also has an marketplace for extensions and JMP also has extensions, one critical feature in statistical computing development is where coders can finally earn some money by creating algorithms, packages and extensions (atleast to compare with game creators on mobiles!!)

 

 

 

Statistical Theory for High Performance Analytics

A thing that strikes me when I was a student of statistics is that most theories of sampling, testing of hypothesis and modeling were built in an age where data was predominantly insufficient, computation was inherently manual and results of tests aimed at large enough differences.

I look now at the explosion of data, at the cloud computing enabled processing power on demand, and competitive dynamics of businesses to venture out my opinion-

1) We now have large , even excess data than we had before for statisticians a generation ago.

2) We now have extremely powerful computing devices, provided we can process our algorithms in parallel.

3) Even a slight uptick in modeling efficiency or mild uptick in business insight can provide huge monetary savings.

Call it High Performance Analytics or Big Data or Cloud Computing- are we sure statisticians are creating enough mathematical theory or are we just taking it easy in our statistics classrooms only to be subjected to something completely different when we hit the analytics workplace.

Do we  need more theorists as well? Is there ANY incentive for corporations with private R and D research teams to share their latest cutting edge theoretical work outside their corporate silo.

 

Related-

“a mathematician is a machine for turning coffee into theorems

On a Hiatus

No blogging (except for interviews)

No poetry (unless I get really inspired and my scrapbook fills up)

No random internet browsing (except search for research)

Hell no, Facebook

No TV

No Movies

No goofing off

No wasting time using creative juice stewing as an excuse

Write the book

Write the book

Write the book, dammit

Stanford Courses Delayed Again

Message from the guys at Palo Alto— Why dont they just make videos using Sal Academy’s help?

We’re sorry to have to tell you that our Machine Learning course will be delayed further. There have naturally been legal and administrative issues to be sorted out in offering Stanford classes freely to the outside world, and it’s just been taking time. We have, however, been able to take advantage of the extra time to debug and improve our course content!

We now expect that the course will start either late in February or early in March. We will let you know as soon as we hear a definite date. We apologize for the lack of communication in recent weeks; we kept hoping we would have a concrete launch date to give you, but that date has kept slipping.

Thanks so much for your patience! We are really sorry for repeatedly making you wait, and for any interference this causes in your schedules. We’re as excited and anxious as you are to get started, and we both look forward to your joining us soon in Machine Learning!

Andrew Ng and the ML Course Staff