Tag: use

Webscraping using iMacros

The noted Diamonds dataset in the ggplot2 package of R is actually culled from the website http://www.diamondse.info/diamond-prices.asp

However it has ~55000 diamonds, while the whole Diamonds search engine has almost ten times that number. Using iMacros – a Google Chrome Plugin, we can scrape that data (or almost any data). The iMacros chrome plugin is available at  https://chrome.google.com/webstore/detail/cplklnmnlbnpmjogncfgfijoopmnlemp while notes on coding are at http://wiki.imacros.net

Imacros makes coding as easy as recording macro and the code is automatcially generated for whatever actions you do. You can set parameters to extract only specific parts of the website, and code can be run into a loop (of 9999 times!)

Here is the iMacros code-Note you need to navigate to the web site http://www.diamondse.info/diamond-prices.asp before running it

VERSION BUILD=5100505 RECORDER=CR
FRAME F=1
SET !EXTRACT_TEST_POPUP NO
SET !ERRORIGNORE YES
TAG POS=6 TYPE=TABLE ATTR=TXT:* EXTRACT=TXT
TAG POS=1 TYPE=DIV ATTR=CLASS:paginate_enabled_next
SAVEAS TYPE=EXTRACT FOLDER=* FILE=test+3

 

 

 

 

 

 

 

 

 

and voila- all the diamonds you need to analyze!

The returning data can be read using the standard delimiter data munging in the language of SAS or R.

More on IMacros from

https://chrome.google.com/webstore/detail/cplklnmnlbnpmjogncfgfijoopmnlemp/details

Description

Automate your web browser. Record and replay repetitious work

If you encounter any problems with iMacros for Chrome, please let us know in our Chrome user forum at http://forum.iopus.com/viewforum.php?f=21

Our forum is also the best place for new feature suggestions :-)
----

iMacros was designed to automate the most repetitious tasks on the web. If there’s an activity you have to do repeatedly, just record it in iMacros. The next time you need to do it, the entire macro will run at the click of a button! With iMacros, you can quickly and easily fill out web forms, remember passwords, create a webmail notifier, and more. You can keep the macros on your computer for your own use, use them within bookmark sync / Xmarks or share them with others by embedding them on your homepage, blog, company Intranet or any social bookmarking service as bookmarklet. The uses are limited only by your imagination!

Popular uses are as web macro recorder, form filler on steroids and highly-secure password manager (256-bit AES encryption).


Interview BigML.com

Here is an interview with Charlie Parker, head of large scale online algorithms at http://bigml.com

Ajay-  Describe your own personal background in scientific computing, and how you came to be involved with machine learning, cloud computing and BigML.com

Charlie- I am a machine learning Ph.D. from Oregon State University. Francisco Martin (our founder and CEO), Adam Ashenfelter (the lead developer on the tree algorithm), and myself were all studying machine learning at OSU around the same time. We all went our separate ways after that.

Francisco started Strands and turned it into a 100+ million dollar company building recommender systems. Adam worked for CleverSet, a probabilistic modeling company that was eventually sold to Cisco, I believe. I worked for several years in the research labs at Eastman Kodak on data mining, text analysis, and computer vision.

When Francisco left Strands to start BigML, he brought in Justin Donaldson who is a brilliant visualization guy from Indiana, and an ex-Googler named Jose Ortega who is responsible for most of our data infrastructure. They pulled in Adam and I a few months later. We also have Poul Petersen, a former Strands employee, who manages our herd of servers. He is a wizard and makes everyone else’s life much easier.

Ajay- You use clojure for the back end of BigML.com .Are there any other languages and packages you are considering? What makes clojure such a good fit for cloud computing ?

Charlie- Clojure is a great language because it offers you all of the benefits of Java (extensive libraries, cross-platform compatibility, easy integration with things like Hadoop, etc.) but has the syntactical elegance of a functional language. This makes our code base small and easy to read as well as powerful.

We’ve had occasional issues with speed, but that just means writing the occasional function or library in Java. As we build towards processing data at the Terabyte level, we’re hoping to create a framework that is language-agnostic to some extent. So if we have some great machine learning code in C, for example, we’ll use Clojure to tie everything together, but the code that does the heavy lifting will still be in C. For the API and Web layers, we use Python and Django, and Justin is a huge fan of HaXe for our visualizations.

 Ajay- Current support is for Decision Trees. When can we see SVM, K Means Clustering and Logit Regression?

Charlie- Right now we’re focused on perfecting our infrastructure and giving you new ways to put data in the system, but expect to see more algorithms appearing in the next few months. We want to make sure they are as beautiful and easy to use as the trees are. Without giving too much away, the first new thing we will probably introduce is an ensemble method of some sort (such as Boosting or Bagging). Clustering is a little further away but we’ll get there soon!

Ajay- How can we use the BigML.com API using R and Python.

Charlie- We have a public github repo for the language bindings. https://github.com/bigmlcom/io Right now, there there are only bash scripts but that should change very soon. The python bindings should be there in a matter of days, and the R bindings in probably a week or two. Clojure and Java bindings should follow shortly after that. We’ll have a blog post about it each time we release a new language binding. http://blog.bigml.com/

Ajay-  How can we predict large numbers of observations using a Model  that has been built and pruned (model scoring)?

Charlie- We are in the process of refactoring our backend right now for better support for batch prediction and model evaluation. This is something that is probably only a few weeks away. Keep your eye on our blog for updates!

Ajay-  How can we export models built in BigML.com for scoring data locally.

Charlie- This is as simple as a call to our API. https://bigml.com/developers/models The call gives you a JSON object representing the tree that is roughly equivalent to a PMML-style representation.

About-

You can read about Charlie Parker at http://www.linkedin.com/pub/charles-parker/11/85b/4b5 and the rest of the BigML team at

https://bigml.com/team

 

Software Review- Google Drive versus Dropbox

Here are some notes from reviewing Google Drive  https://drive.google.com/ vs Dropbox https://www.dropbox.com/.

1) Google Drive gives more free space upfront  than Dropbox.5GB versus 2GB

2) Dropbox has a referral system 500 mb per referral while there is no referral system for Google Drive

3) The sync facility with Google Docs makes Google Drive especially useful for prior users of Google Docs.

4) API access to Google Drive is only for Chrome apps which is intriguing!

https://developers.google.com/drive/apps_overview

Apps will not have any API access to files unless users have first installed the app in Chrome Web Store.

You can use the Dropbox API much more easily –

See the platforms at

https://www.dropbox.com/developers/start/core

Choose your platform:

iOS Android Python Ruby

But-

(though I wonder if you set the R working directory to the local shared drive for Google Drive it should sync up as well but of course be slower –http://scrogster.wordpress.com/2011/01/29/using-dropbox-with-r-2/)

5) Google Drive icon is ugly (seriously, dude!) , but the features in the Windows app is just the same as the Dropbox App. Too similar 😉

 

6) Upgrade space is much more cheaper to Google Drive than Dropbox ( by Google Drive prices being exactly  a quarter of prices on Dropbox and max storage being 16 times as much). This will affect power storage users. I expect to see some slowdown in Dropbox new business unless G Drive has outage (like Gmail) . Existing users at Dropbox probably wont shift for the small dollar amount- though it is quite easy to do so.

 

Install Google Drive on your local workstation and cut and paste your Dropbox local folder to the Google Drive local folder!!

7) Dropbox deserves credit for being first (like Hotmail and AOL) but Google Drive is almost better in all respects!

Google Drive

Free
5 GB of Drive (0% used)
10 GB of Gmail (48% used)
1 GB of Picasa (0% used)

Upgrade:

25 GB
2,49 $ / Month
+25 GB for Drive and Picasa
Bonus: Your Gmail storage will be upgraded to 25 GB.
Choose this plan

100 GB
4,99 $ / Month
+100 GB for Drive and Picasa
Bonus: Your Gmail storage will be upgraded to 25 GB.
Choose this plan

 Need more storage?

Up to 16 TB available

Dropbox–

Current account type

Large DropboxDropbox Badge greenFree
Free
Up to 18 GB (2 GB + 500 MB per referral)
 Account info 

Other account types

Large DropboxDropbox Badge orange50 GB +
Pro 50
+1 GB per referral, up to +32 GB
 $9.99/month or $99.00/year Upgrade to Pro 50
Large DropboxDropbox Badge purple100 GB +
Pro 100
+1 GB per referral, up to +32 GB
 $19.99/month or $199.00/year Upgrade to Pro 100
Triple DropboxDropbox For Teams Badge1 TB +
Teams
Plans starting at 1 TB
 Large shared quota, centralized admin and billing, and more!

 

 

 

Software Review- BigML.com – Machine Learning meets the Cloud

I had a chance to dekko the new startup BigML https://bigml.com/ and was suitably impressed by the briefing and my own puttering around the site. Here is my review-

1) The website is very intutively designed- You can create a dataset from an uploaded file in one click and you can create a Decision Tree model in one click as well. I wish other cloud computing websites like  Google Prediction API make design so intutive and easy to understand. Also unlike Google Prediction API, the models are not black box models, but have a description which can be understood.

2) It includes some well known data sources for people trying it out. They were kind enough to offer 5 invite codes for readers of Decisionstats ( if you want to check it yourself, use the codes below the post, note they are one time only , so the first five get the invites.

BigML is still invite only but plan to get into open release soon.

3) Data Sources can only be by uploading files (csv) but they plan to change this hopefully to get data from buckets (s3? or Google?) and from URLs.

4) The one click operation to convert data source into a dataset shows a histogram (distribution) of individual variables.The back end is clojure , because the team explained it made the easiest sense and fit with Java. The good news (?) is you would never see the clojure code at the back end. You can read about it from http://clojure.org/

As cloud computing takes off (someday) I expect clojure popularity to take off as well.

Clojure is a dynamic programming language that targets the Java Virtual Machine (and the CLR, and JavaScript). It is designed to be a general-purpose language, combining the approachability and interactive development of a scripting language with an efficient and robust infrastructure for multithreaded programming. Clojure is a compiled language – it compiles directly to JVM bytecode, yet remains completely dynamic. Every feature supported by Clojure is supported at runtime. Clojure provides easy access to the Java frameworks, with optional type hints and type inference, to ensure that calls to Java can avoid reflection.

Clojure is a dialect of Lisp

 

5) As of now decision trees is the only distributed algol, but they expect to roll out other machine learning stuff soon. Hopefully this includes regression (as logit and linear) and k means clustering. The trees are created and pruned in real time which gives a slightly animated (and impressive effect). and yes model building is an one click operation.

The real time -live pruning is really impressive and I wonder why /how it can ever be replicated in other software based on desktop, because of the sheer interactive nature.

 

Making the model is just half the work. Creating predictions and scoring the model is what is really the money-earner. It is one click and customization is quite intuitive. It is not quite PMML compliant yet so I hope some Zemanta like functionality can be added so huge amounts of models can be applied to predictions or score data in real time.

 

If you are a developer/data hacker, you should check out this section too- it is quite impressive that the designers of BigML have planned for API access so early.

https://bigml.com/developers

BigML.io gives you:

  • Secure programmatic access to all your BigML resources.
  • Fully white-box access to your datasets and models.
  • Asynchronous creation of datasets and models.
  • Near real-time predictions.

 

Note: For your convenience, some of the snippets below include your real username and API key.

Please keep them secret.

REST API

BigML.io conforms to the design principles of Representational State Transfer (REST). BigML.io is enterely HTTP-based.

BigML.io gives you access to four basic resources: Source, Dataset, Model and Prediction. You cancreate, read, update, and delete resources using the respective standard HTTP methods: POST, GET,PUT and DELETE.

All communication with BigML.io is JSON formatted except for source creation. Source creation is handled with a HTTP PUT using the “multipart/form-data” content-type

HTTPS

All access to BigML.io must be performed over HTTPS

and https://bigml.com/developers/quick_start ( In think an R package which uses JSON ,RCurl  would further help in enhancing ease of usage).

 

Summary-

Overall a welcome addition to make software in the real of cloud computing and statistical computation/business analytics both easy to use and easy to deploy with fail safe mechanisms built in.

Check out https://bigml.com/ for yourself to see.

The invite codes are here -one time use only- first five get the invites- so click and try your luck, machine learning on the cloud.

If you dont get an invite (or it is already used, just leave your email there and wait a couple of days to get approval)

  1. https://bigml.com/accounts/register/?code=E1FE7
  2. https://bigml.com/accounts/register/?code=09991
  3. https://bigml.com/accounts/register/?code=5367D
  4. https://bigml.com/accounts/register/?code=76EEF
  5. https://bigml.com/accounts/register/?code=742FD

Oracle R Updated!

Interesting message from https://blogs.oracle.com/R/ the latest R blog

 

_——–_

Oracle just released the latest update to Oracle R Enterprise, version 1.1. This release includes the Oracle R Distribution (based on open source R, version 2.13.2), an improved server installation, and much more.  The key new features include:

  • Extended Server Support: New support for Windows 32 and 64-bit server components, as well as continuing support for Linux 64-bit server components
  • Improved Installation: Linux 64-bit server installation now provides robust status updates and prerequisite checks
  • Performance Improvements: Improved performance for embedded R script execution calculations

In addition, the updated ROracle package, which is used with Oracle R Enterprise, now reads date data by conversion to character strings.

We encourage you download Oracle software for evaluation from the Oracle Technology Network. See these links for R-related software: Oracle R Distribution, Oracle R Enterprise, ROracle, Oracle R Connector for Hadoop.  As always, we welcome comments and questions on the Oracle R Forum.

 

 

Oracle R Distribution 2-13.2 Update Available

Oracle has released an update to the Oracle R Distribution, an Oracle-supported distribution of open source R. Oracle R Distribution 2-13.2 now contains the ability to dynamically link the following libraries on both Windows and Linux:

  • The Intel Math Kernel Library (MKL) on Intel chips
  • The AMD Core Math Library (ACML) on AMD chips

 

To take advantage of the performance enhancements provided by Intel MKL or AMD ACML in Oracle R Distribution, simply add the MKL or ACML shared library directory to the LD_LIBRARY_PATH system environment variable. This automatically enables MKL or ACML to make use of all available processors, vastly speeding up linear algebra computations and eliminating the need to recompile R.  Even on a single core, the optimized algorithms in the Intel MKL libraries are faster than using R’s standard BLAS library.

Open-source R is linked to NetLib’s BLAS libraries, but they are not multi-threaded and only use one core. While R’s internal BLAS are efficient for most computations, it’s possible to recompile R to link to a different, multi-threaded BLAS library to improve performance on eligible calculations. Compiling and linking to R yourself can be involved, but for many, the significantly improved calculation speed justifies the effort. Oracle R Distribution notably simplifies the process of using external math libraries by enabling R to auto-load MKL orACML. For R commands that don’t link to BLAS code, taking advantage of database parallelism usingembedded R execution in Oracle R Enterprise is the route to improved performance.

For more information about rebuilding R with different BLAS libraries, see the linear algebra section in the R Installation and Administration manual. As always, the Oracle R Distribution is available as a free download to anyone. Questions and comments are welcome on the Oracle R Forum.

New Economics Theories for the new Tech World

When I was doing my MBA (a decade ago), one of the principal theories on why corporations exist was 1) Shareholder Value creation (grow wealth for investors) and a notable second was 2) Stakeholder Value creation- creating jobs for societies, providing tax to countries, providing employees with stable employment and incentives,  and of course creating monetary value for shareholders.

There were two ways you could raise money- debt or equity. Debt had the advantage of interest payments being tax deductible. Debt payments had to be met regularly. Equity had the advantage that equity holders were the last ones to be paid in case of closing the company down, which justified that rate of return on equity is generally higher than cost of debt.  Dividend payouts to stockholders could be deferred in a low revenue year or due to planning reasons.

Or in plain English, over the long term borrowing money from share holders in lieu of stocks was more expensive than selling bonds or borrowing from the banks.

Hybrid combinations of debt and equity were warrants and debentures that started off as one form of instrument and over a period of time gave much more flexibility and risk safety nets to both issuers and subscribers of capital. Another hybrid was stock options (now considered as a default option of rewarding employees in technology companies, but this was not always the case).

The use of call and put options in debentures, and the idea of vesting period in stock options was to promote lone term stability and minimize fluctuations in stock prices, employee attrition, besides of course to minimize the weighted average cost of capital. Venture capital was another class of capital known for both huge rates of return and risk taking (?)

But in today’s world where a Google has three classes of shares, companies trade shares before IPOs, and valuations of technology companies sink and rise by huge % over weeks (especially as they near IPO dates)- I wonder if traditional theories in finance need a much stronger overhaul.

or do markets need a regulatory overhaul, that would enable stock exchanges to have once more the credibility they had as the primary sources of raising capital.

 

Who will guard the guardians? Their conscience- the regulators or the news media?

There are ways of raising money that are not evil.

But they are not perfectly fair as well.

Color Palettes in R using RColorBrewer #rstats

The lovely colors at http://ColorBrewer.org can be used for much better color palettes in R.

library(RColorBrewer)

display.brewer.all()

and we use the function brewer.pal(N,”Name”) as the col  parameter for the new color palettes

where we can see name of palettes  from the list above

data(VADeaths)
par(mfrow=c(2,3))
 hist(VADeaths,col=brewer.pal(3,"Set3"),main="Set3 3 colors")
 hist(VADeaths,col=brewer.pal(3,"Set2"),main="Set2 3 colors")
 hist(VADeaths,col=brewer.pal(3,"Set1"),main="Set1 3 colors")
 hist(VADeaths,col=brewer.pal(8,"Set3"),main="Set3 8 colors")
 hist(VADeaths,col=brewer.pal(8,"Greys"),main="Greys 8 colors")
 hist(VADeaths,col=brewer.pal(8,"Greens"),main="Greens 8 colors")
Created by Pretty R at inside-R.org

Rplot7

• Colors from [http://www.ColorBrewer.org] by Cynthia A. Brewer, Geography, Pennsylvania State University
• Erich Neuwirth (2011). RColorBrewer: ColorBrewer palettes. R package version 1.0-5. [http://CRAN.R-project.org/package=RColorBrewer]
Note-ColorBrewer is Copyright (c) 2002 Cynthia Brewer, Mark Harrower, and The Pennsylvania State University. All rights reserved. The ColorBrewer palettes have been included in the R package with permission of the copyright holder.