Big Data: Four Factors to Tune your Meaningful Analytics

The last two weeks I have talked about ways to choose meaningful business analytics and then fine tune the focus of your business analytics. This week I talk about the factors needed to further fine tune your analytics into effective results for your company.   Some of the ideas might even help you get the best team picks for the NCAA Warren Buffet $1 Billion bet for the prefect brackets challenge.  Listening to Tim Chartier from Davidson College in this Bloomberg interview as he discusses using analytics to predict NCAA winners with a 90%+ success rate.  His 90%+ success rate highlights the following three factors you can use to fine tune your own company’s analytics.

1. Determine what specific analytics impact your goals.  – Factors such as offensive scoring, defense, coaching, and physical player attributes help determine who will win a basketball game.  Within the model Tim Chartier has developed, evaluated, and weighted over several years, there are 350 different factors used to predict a basketball game outcome.  Your analytics teams  and business subject matter experts need to have discussions to help determine your specific analytical factors to provide a winning business result.  How many analytics aspects are necessary for your team to develop and how long will it take to develop the multitudes of different factors required for your company’s goal or situation?
2. Be able to add new factors to improve your analytics results.  In his Bloomberg interview Tim Chartier says that coaching is not factored into his NCAA predictions, but it is another factor his analytics team is looking to integrate into its analysis.  Your analytics effort will be a learning and evaluation process to achieve your sales or bottom line goal.  Management must be flexible enough to understand that learning will occur over time. Analyzing the correct factors will provide correct results, but adding new factors for a deeper understanding is always necessary.  Integrating new factors into your business analytics efforts should be easy.  Make sure that your processes include procedures that can easily incorporate new factors into your analytics so that your success percentage can evolve and improve.  
   
   
3. Experiment with different weighting of your analytic factors.  By using more than 350 NCAA team factors, Tim Chartier’s analytical process has a greater than 90% success rate. The analytical process uses different factor weightings to fine tune the prediction success.  There are many different factors and predictive analytic modeling types.  Since there are so many different analytic model types, your company’s analytics is going to be a diverse set of analytics that are unique to your market, products, and location.  Make sure to design your analytics procedures to be able to change the weightings of different factors and compare the results to your previous outcomes in order to fine tune your analytics to the best they can be.  
   
   
4. You don’t necessarily need big data. In this article about Kaggle’s data sharing and analytics processes, William Cukierski comments that the big data analytics phase the industry is in is over hyped.  Your analytics doesn’t necessarily need petabytes of information to get to a 90%+ success factor.  The article discusses analytics using elimination of bad lima beans on a conveyer belt as an example.  If you have seen five examples of bad brown lima beans, it’s easy to pick out other bad brown lima beans.  Think about the uniqueness of analytics needed for your project and cut your big data requirements down to size for your company situation.

Analytics is being developed against all type of medical, social, government, and other industries to gain understanding to make things better.  Use the ideas listed over the past few weeks to get the best out of your analytics. I hope you win the billion in the NCAA contest if I don’t.
_______________________________________________
Dave Beulke is a system strategist, application architect, and performance expert specializing in Big Data, data warehouses, and high performance internet business solutions. He is an IBM Gold Consultant, Information Champion, and President of DAMA-NCR, former President of International DB2 User Group, and frequent speaker at national and international conferences. His architectures, designs, and performance tuning techniques help organization better leverage their information assets, saving millions in processing costs.

Don’t forget to make your reservations for the IDUG DB2 Technical Conference in Phoenix, Arizona this May 12-16th.  Also the IDUG European conference is still accepting presentation abstracts to be potentially pick for the IDUG EU conference in Prague, Czech Republic November 9-14th.  For more details go to www.idug.org.  Also plan on attending my presentation “Big Data Disaster Recovery Performance” at the IDUG conference in Phoenix.