Thursday, May 16, 2013

Live from FOCI 2013: Big Data: Powerful Predictions Through Data Analytics

Nate Silver, a world renowned statistician and founder of fivethiryeight.com, spoke of the world of errors and predictions, very relevant to the big data environment. He evoked the thought of what kind of predictions we can trust, and how much can we trust forecasters?

Judging from Hurrican eSandy or terror attacks or unforgiving hacked tweets and the widespread reach of all these, the topic is very relevant in a world where we want to know what happens before it happens, and wish to micromanage while it happens.

Nate's 4 suggestions are as follows:

1. Think Probabalistically: convey uncertainty knowing what can go wrong. Only if you know what you do can you know what goes wrong.

2. Know Where You're Coming From: Know that you have a point of view, it is helpful in identification and forecasting.

3. Survey the Data Landscape: What makes data rich? Quality, quantity and variety. Understanding if you're in a data rich or data poor environment is critical.



4. Try and Err: Experiment with real data, test hypotheses, segment the big data, and converge to a great solution.

We live in a big data and uncertain world. Never has it been more obvious that we need analytics to navigate better.



Sourabh Sharma, Communication & Social Media Research Expert at SKIM, an international consultancy and marketing research agency, has a background in engineering, marketing and finance from the University of Pennsylvania, and the Wharton School and Rotterdam School of Management. Having worked in marketing and product development at L’Oreal, followed by a stint in management consulting, he now passionately enjoys the world of social media, and can be found on every platform with his alias sssourabh. He is a food critic and a fashion writer, and documents these alongside strategy on his blog called 3FS. He may be reached at s.sharma@skimgroup.com. Follow him on @sssourabh.


No comments: