Could data mining techniques accurately predict the medal counts at the Olympics?  This question came into mind four years ago while watching the Winter Games in Vancouver.  I knew that a predictive model could give us an estimate of the number of medals each nation might win; but how close could we get to the actual outcomes? It was a tantalizing project ….          …

Continue

When creating a predictive model, data miners need to “tune” it to make the right kind of mistakes.  Setting the cut-off point between ‘promising’ and ‘unpromising’ depends a lot on our client’s biggest concern -- missed opportunities or false alarms.

Read the full…

Continue

Improving our presentation skills will make for happier clients and bosses.  In this article, Tim Graettinger focuses on five crucial practices that will make your next presentation twice as good as your last.…

Continue

Improving our presentation skills will make for happier clients and bosses.  In this article, Tim Graettinger focuses on five crucial practices that will make your next presentation twice as good as your last.

Read the…

Continue

Is it possible to use data mining techniques to predict how many medals each nation will win at the Olympic Games?  What is it about a nation that allows it to produce Olympic medal-winning athletes?  Why do some countries take home a bundle of medals while others don't?  As we think more and more about it, it quickly becomes clear that the why’s behind the medal counts at the Olympics are complex.…

Continue

In this article, Tim Graettinger focuses on three key questions for clearly and explicitly defining the problem to be solved at the outset of a data mining project.

• What is the outcome?
• What is the time frame?
• How will we measure success?

Read the article…

Continue

For all the strides that data mining tools have made, using them well still requires hard work and critical thought.  In this article on the nuts & bolts of DM, we review a real workhorse for data mining and analysis - the histogram.  Among the histograms encountered most frequently in practice are the following: “money”, “count”, and “outlier”.  We will look at each one of them in turn….     Read the…

Continue

In this article, we review two, intertwined tools in the data mining arsenal: correlation and scatter plots.  These tools are fundamental for gauging the relationship (if any) between pairs of data elements (say, the age and income of your customers).  As we’ll soon see, there are good, bad, and ugly things that can happen when we apply a purely computational method like correlation.  My goal is to help you avoid the usual pitfalls, so  you can use correlation and scatter plots effectively…

Continue

In the data mining process, where do data scientists like you and me add the most value?  Is it in exploring the data, uncovering anomalies and seeing relationships between elements?  In selecting transformations for the elements to improve their representations for modeling and analysis?  In building sophisticated predictive models?  For my money, the answer is ‘none of…

Continue

Many customer behaviors have the flavor of a choice between two alternatives:  Yes or no.  Buy or sell.  Renew or cancel.  Suppose software called a “classifier” is available to predict customer choices in advance.  Would you use it?  Perhaps you’d like to test it to see how well it performs before you commit.  In this installment of my series on the nuts and bolts of data mining, I discuss the use of classifiers and questions about their performance.  Regarding performance, we specifically…

Continue

Another recent article from Discovery Corps, Inc. president Tim Graettinger:   Grab Bag 2:  More Frequently-Asked Questions about Data Mining.  

"While helping present a monthly webinar on Data Mining, I'm asked some challenging and really pivotal questions about DM and predictive analytics:  

1. How much data do I need? 
2. Why doesn't my predictive model perform as well on new data as it did on the training data?
3. Are new DM algorithms needed to…

Continue

My company (Discovery Corps, Inc.)  president Tim Graettinger has released a new article I think you'll like:

"If you’re a data miner, there’s a good chance you saw the movie Moneyball.  For me, the movie was entertaining, but the book was really enlightening.  I found it to be a source of numerous and familiar lessons about data mining - while couched in the intriguing context of major league baseball.  In this article, I’d like to share some of…

Continue

As a little diversion this summer, we at Discovery Corps decided to see if we could predict the finalists on 'Dancing with the Stars' even before the contestants take to the dance floor.  Data analytics professionals know that statistics can be fun, so we dove into it.  We got some really interesting results.  Have a look at the article on our website.

 …

Continue

An article on INTERVIEWING SKILLS:  I recently posted another article on my company's website, this one on interviewing skills for data miners and modelers.  Take a look, and see what companies are looking for in data mining candidates.  So You Want a Job: Interviewing Data Miners

I thought I'd post another article from my company's president, Tim Graettinger.  Here's a little intro:  "More and more individuals and firms are taking steps to reduce their carbon footprints.  What about data mining - can it contribute to a company's efforts to have a positive impact on the environment?  My answer is an unequivocal 'yes', and in this article you'll learn the reasons why.  We'll explore a sampler platter of data mining applications that can benefit you, your company, your… Continue

The president of the data mining company I work for recently published a new article on data mining.  This article from Tim Graettinger addresses some of the top questions he's been asked during a webinar on data mining that he helps to present:

• What tools or tool sets does he recommend?
• How can I get management to 'buy in' on a data mining project?
• How can I transform non-numeric data so I can use it easily in predictive modeling?

His responses… Continue