Friday, March 16, 2012

That's a Slam Dunk: Basketball Analytics

It turns out that moneyball, the technique (book and movie) of applying analytics to baseball, is also being applied to the NBA.  A recent conference at MIT looked at the various ways in which basketball can be understood by applying the analytics lens.  Some of the fascinating paper topics included a discussion about whether playoff experience matters in winning championships, and how to calculate the odds of choking a free throw.  Clearly,  the business of professional sports is ripe for use of analytics.  Now if someone would just figure out a way to use analytics to predict the NCAA tournament winner...

Read the article.

(Submitted by Bharath)

Deconstructing the Galactic Empire for Leadership Best Practices

Alex Knapp of Forbes.com penned this very clever article analyzing the five leadership mistakes made by the Galactic Empire in the Star Wars movies (IV-VI). Interestingly, many of the mistakes highlighted here also apply to high tech companies.  For example, Knapp cites "focusing all of the organization’s efforts into a single goal and failing to consider alternatives" as a key takeaway.  Not only can that attitude apply to many tech startups, but it typified the dot com era.  Whether you are looking for other leadership pearls of wisdom, or are just a Star Wars fan, this is a great read.  If you happen to be a Star Trek fan, there are links to leadership lessons from that series posted here as well.

Read the article.

(submitted by Robin)

No Email For You (after hours)

Volkswagen is bucking the trend when it comes to work-based communication.  The company turns off their Blackberry servers after hours to ensure employees have a better work-life balance.  This practice is a shift away from the norm of 24X7 connectivity, but it certainly has its merits.  According to the article emails stop 30 minutes after a shift ending and start 30 minutes prior to a shift beginning.  Apparently, the rule does not apply to senior management.  There are a number of cultural, legal, and business factors that go into a decision like this, so it may not be the right one for all (or even any other) companies, but it is an interesting decision that deserves a closer look.

Read the article.

Is Google+ a Bust?

Based on pure metrics, it appears so.  Google+ only has 100 million users that spend a couple minutes per month on the site compared to Facebook's 850 million users who are logged in for nearly eight hours per month.  To be fair, Google+ has only been around for six months, but usage seems to have already plateaued.  However, according to this article and tech evangelist Guy Kawasaki, Google+ may be a longer-term play and comparing it to other social networks may be a little misleading.  Google+ is more of a social layer that is applied to the various Google products, making it a very different value proposition;  Google+ is a collection of activity done on various Google sites, instead of a site that people connect to like Facebook, Twitter, etc.  In this way, even though short-term success has been limited, there may be significant long-term success to come.


Read the article.

(submitted by Tom)

Yes We Can! Predicting Elections Through Social Media

What is the best way to predict the results of the upcoming elections? How about listening to what people are saying? Election Predictors is mining the text from social media sites to determine the mood of potential voters. This article talks of predictions by Attensity, a text mining company, that Mitt Romney would win seven states on Super Tuesday based on more than 800,000 posts on Twitter.  Four of the predictions were predicted correctly (MA, VA, ID, and OH). Forrester and Carnegie Melon researchers stress that this technique is still in its infancy and that more progress needs to be made in terms of weeding out artificial content and detecting sarcasm to ensure greater accuracy of predictions.  Also, adding a larger data set, like Facebook, could make this process far more accurate.  In a business context, companies looking to maximize social media presence may want to go beyond handling customer complaints on Twitter and develop adaptive product and service offerings based on the indirect customer feedback freely available.

Read the article.

(submitted and written by Kyle)

Friday, March 9, 2012

Using Predictive Analytics and Cell Phone Usage to Build a Credit Score


This article describes how one company has created software that analyzes a customer's cell phone usage data to create a lifestyle profile and apply a credit rating score.  The software factors in "length of calls, time of day, and location."  For example, a user who makes phone calls in the middle of the night on a prepaid phone card could be considered a credit risk. Essentially, every phone call provides a set of data points that can be incorporated into building these complex predictive models.  Single tier data mining (of cell phone records in this case) is nothing new, but worth noting here is how the mined data is extrapolated for use in another domain, such as creating a credit rating score.

Read the article.

How Secure is Your password, Password1, Password01?

The Trustwave security consulting firm just released a report summarizing the results of 2 million network vulnerability scans and 300 security breach investigations.  It turns out that most hacking incidents occur because of weak passwords, not sophisticated hacker tools. In their research they found that 5% of passwords contain some variation of the word "password."

This report illustrates a basic problem with username/password based security- its vulnerability to the limitations of the end user.  Most users choose passwords that are too weak and easily guessable.  To combat this, user management tools usually require users to include in their password at least one capital letter and a mix of letters and numbers.  Naturally, users create words they can remember like Password1.  Hackers are able to run brute force hacking attacks by running a program that attempts logins with variations of words from a dictionary (called a dictionary attack).  The tool will try password, Password, Password1, Password2, etc...  With the massive amount of processing power in most PCs these days, hackers can run millions of password guessing permutations in minutes.  The more-secure alternative is to use a system-generated strong password that is less vulnerable to a dictionary attack, but these are often complex and hard to remember, prompting users to write them down and store them where they can be stolen.

The digital universe is rapidly getting to the point where relying solely on username/password security is inadequate.  Either new technology methods will have to be developed or systems and providers will have to embrace multi-factor authentication (MFA).  MFA users must validate themselves with more than one authentication method.  These modes are defined as something you know (password), something you have (digital access card/token), or something you are (finger/handprint, retina scan).  While it is hard to imagine widespread adoption of MFA, it is also shocking to imagine that more places haven't required these security enhancements, given the weaknesses of user-generated passwords.  This article should be a call to action- let's see who responds to it.

Read the article.

Exploring Hadoop, Without the Hype

Hadoop is one of the hottest buzzwords flying around these days, although most people don't know what it is.  In simple terms, Hadoop is an open-source distributed data processing platform.  This underlying architecture allows for large-scale data processing tasks to be broken down into chunks and distributed across multiple nodes. These nodes can be cheap low-cost clustered servers that create a cost-effective distributed platform.  It makes use of a custom parallel processing language called Pig.  Hadoop's most notable feature is its ability to process unstructured data (although it can process structured data types as well).  The big analytics vendors (IBM, HP, EMC, and Microsoft) have all jumped in to create commercial versions of Hadoop, so it's safe to say we can expect big things (as well as a lot of hype) from Hadoop in the future.

If you have an interest in Big Data, this article is a must read.  It provides a deep dive understanding of what Hadoop is and how it can be effectively applied.

Read the article.

(submitted by Judi)

It's Not Just About the new iPad: 315 Million Devices Can't be Wrong

There's a lot of excitement surrounding Apple's iPad3: the new display, connection to faster cellular networks, and quad-core chip make it a very compelling device.  Undoubtedly, the device will sell well and should continue to dominate the tablet market for some time.  But something more profound (though admittedly less sexy) emerged from Apple CEO Tim Cook's keynote announcement this week.  Cook noted that Apple has sold 315 million iOS devices (iPad, iPhone, iPod) to date, including 62 million units in Q4 of 2011 alone. This underscores a trend, dubbed "The Post-PC Revolution," that analysts have been tracking for awhile - the cannibalizing of the PC market by smartphones and tablet devices.  Apple's numbers testify to the fact that while PCs are not going away, they certainly are losing their dominance, market share, and to some degree, innovative development. This paradigm shift will have profound implications for longer-term IT strategies and will affect everything from security to procurement.

Read the article.

What the Heck is a Data Scientist?

The term "data scientist" is bandied about almost as much as "Big Data" these days.  But what exactly does this term mean and why should we care?  It is an umbrella term that broadly refers to people working in the advanced analytics field, exploring ways in which data can be used to transform a business, government, or even the world.  Data scientists seek to use data in new ways and/or re-conceptualize what it means to use data.  In the process they develop new tools, processes, and even definitions.  The article below does a good job laying out a practical understanding of this term.

Read the article.

(submitted by Bharath)