Rogue Wolves

Language Performance Benchmarks

2012-12-12T00:00:00-08:00

Recently, I came across Julia, a high-level, high-performance dynamic programming language for technical computing. The language itself is interesting, but on the Julia website is are comparative language performance benchmark results. Julia is compared against several other common scientific computing languages, such as, Python, Matlab, R, etc.

The following micro-benchmark results are from a MacBook Pro with a 2.53GHz Intel Core 2 Duo CPU and 8GB of 1066MHz DDR3 RAM

(Figure from julialang.org)

While the performance benchmarks are touting the advantages of the Julia language, I found it most striking the JavaScript performance. It was a surprise to me that JavaScript mostly outperforms Python by an order of magnitude. I knew the optimizations Google and Apple were performing made dramatic improvements, but putting it into this perspective is very impressive. It would have more interesting if they had included Java in the benchmark as well.

I guess I see why node.js has been getting lots of attention for server side processing.

Rice University Data Visualization Course Materials

2012-12-11T00:00:00-08:00

Looking to learn more about data visualization? Hadley Wickham at Rice University has posted all the course materials for his course Stat 645: Data Visualization online. It’s a nice collection of papers, tutorials and presentations covering a wide range of topics related to visualization and visual analytics.

Fix Eclipse Fonts on Retina MacBook Pro

2012-12-08T00:00:00-08:00

Are you being driven mad by the horrible non-retina font rendering of Eclipse on a retina MacBook Pro? Fortunately, there is a workaround:

Open Finder and right click on the Eclipse.app then select “Show package contents”. Edit Contents/Info.plist file in a text editor and add lines 5 + 6.

1     <dict>
2     <plist>
3     ...
4 
5     <key>NSHighResolutionCapable</key>
6     <true/>
7 
8     </dict>
9     </plist>

Then, log out or make a copy of the app so that OSX will notice the change. Launch Eclipse and it should now be retina.

Rapid Machine Learning with Ramp

2012-11-25T00:00:00-08:00

Ramp is a python library for rapid machine learning prototyping. It provides a simple, declarative syntax for exploring features, algorithms and transformations quickly and efficiently. The library leverages pandas, which provides high-performance, easy-to-use data structures and data analysis tools, as well as various python machine learning and statistics libraries (scikit-learn, rpy2, etc.). Some features:

Fast caching and persistence of all intermediate and final calculations – nothing is recomputed unnecessarily.
Advanced training and preparation logic. Ramp respects the current training set, even when using complex trained features and blended predictions, and also tracks the given preparation set (the x values used in feature preparation – e.g. the mean and stdev used for feature normalization.)
A growing library of feature transformations, metrics and estimators. Ramp’s simple API allows for easy extension.

kvh has a nice blog post introducing Ramp. I’m looking forward to experimenting with it.

Script Excel with Python using IronSpread

2012-06-11T00:00:00-07:00

IronSpread is an Excel 2010 plugin for Windows that lets you script your spreadsheets with Python.

I’d choose python over VBA any day.

The Minds of Modern Mathematics App

2012-04-08T00:00:00-07:00

IBM has released a new iOS app, The Minds of Modern Mathematics, which provides an interactive history of mathematics. The app recreates a 50-foot long info graphic designed by Charles and Ray Rames that depicts 1000 years of the history and influence of mathematics through the ages.

If you have ever wondered how mathematicians and mathematics have shaped the world we live in, this is the app for you. The Minds of Modern Mathematics takes you on an interactive journey that spans nearly 1,000 years. It tells the story of mathematics and how it has impacted almost every aspect of human progress, from science to music, art, architecture and culture.

Learning from Data Course

2012-03-25T00:00:00-07:00

Yet another free online introductory course on Machine Learning. This one from Caltech.

The course description from the web page reads:

This is an introductory course on machine learning that covers the basic theory, algorithms, and applications. Machine learning (ML) enables computational systems to adaptively improve their performance with experience accumulated from the observed data. ML techniques are widely applied in engineering, science, finance, and commerce to build systems for which we do not have full mathematical specification (and that covers a lot of systems). The course balances theory and practice, and covers the mathematical as well as the heuristic aspects. Detailed topics are listed below.

I like the (apparent) dig at another free online ML course:

A real Caltech course, not a watered-down version

Ah rivalries…

It really is astonishing the amount and quality of free online education available today. There had never been a better time to sharpen your skills or learn something new.

Introduction to Data Science course

2012-03-25T00:00:00-07:00

Interesting introductory course offered at Berkeley on Data Science. From the course page:

Organizations use their data for decision support and to build data-intensive products and services. The collection of skills required by organizations to support these functions has been grouped under the term “Data Science”. This course will attempt to articulate the expected output of Data Scientists and then equip the students with the ability to deliver against these expectations.

I’m still not a fan of the marketing term “data science”, however more applied machine learning, statistical analysis and data mining courses are definitely a good thing.

Impending Windows 8 Disaster

2012-03-14T00:00:00-07:00

Interesting video from Chris Pirillo posted on the Fixing Windows 8 blog. The video illustrates (quite effectively) the major usability issues with Microsofts new operating system. In the video, we see Chris’ dad trying to navigate between the Metro/Windows modes. It’s painful.

I noticed this problem when some of my colleagues returned from a Windows Developer Conference. They demoed a pre-release version of Windows 8. What struck me about Metro and the multi-modal environment was the expansive use of “magical gestures” that were used pervasively throughout. I made a comment that this was a major UX problem. There is no discoverability of these features. You just have to know a slew of new unintuitive gestures to interact with the operating system.

Since seeing this internal demo, my prediction has been: Windows 8 is going to be an utter disaster for the company. I think it will be on par - if not worse - than the Vista disaster of 2007. I hope I’m wrong, but unless they fix these UX problems post-haste, it’s going to be rough going.

This is a big gamble for Microsoft, but I have little confidence it will pay off. Metro is an interesting and fresh innovation, but I’m unconvinced it will have appeal outside of phones or tablets. Bolting it on as a separate side-car to a “Metro-ish” version of Windows is muddled and confused. Further confusing, is the apparent need to switch back and forth between these modes and lack of integration between them.

Cassovary Large Scale Graph Processing

2012-03-11T00:00:00-08:00

Twitter recently released an open source large scale graph precessing library named Cassovary. The library is written in Scala and therefore can be easily integrated into any JVM-based language. The library focus is on simplicity and efficient data structures to support processing of large graphical structures.

From the project README:

Cassovary is a simple “big graph” processing library for the JVM. Most JVM-hosted graph libraries are flexible but not space efficient. Cassovary is designed from the ground up to first be able to efficiently handle graphs with billions of nodes and edges. A typical example usage is to do large scale graph mining and analysis of a big network.

The library appears to mainly provide data structures and some common graph search algorithms. Any dataming or graph related operations are currently not part of the library. It looks interesting and should generate a lot of interest.