Analytics and US Presidential Elections!

Have you observed the statistics or mathematicians views on Big Data or Analytics?  They probably wouldn’t stress those jargons too much in their conversations.   Checkout Political Statistician Nate Silver’s interview and a question on Big Data.

In the world of Enterprise IT, there is no dearth of jargons – Business Intelligence, Analytics, Big Data, Insights, Data Mining, etc. There is good amount of time one needs to spend NOT on applying these techniques, but to get a perspective on what is real and what is not.  In Nate Silver’s words, in analytics world, one needs to have the ability to differentiate noise from the real signal, ability to differentiate real data from insignificant information.  I believe, in the World of Big Data, its lot more critical to understand and differentiate noise from the signal. Else, we could be easily misled in our decisions.

In the upcoming US presidential elections, Nate Silver predicts President Obama has a 73.6% chance of victory. Yes, 'Chance’ of victory, even though the number 73 is far more significant. Nate Silver humbly claims that he is by no means certain and that his analysis is a mere probability of a given result.  As this article points out, If a Weather forecaster tells you there’s an 80.9% chance of rain, she is not guaranteeing rain. But she is saying you should probably take an umbrella when you step out.

That’s exactly the perspective we should take when we approach Analytics in Enterprise IT – As Nate puts it, any model of real world is an approximation.  He also adds Data driven predictions can succeed or fail. Its when we deny “our role” in the process that the chances of failure rise.  To me, “our role” indicates the magnitude of organizational changes and people behaviors required to adopt analytics successfully.  We discussed about it in my previous post – Big Fat Truth behind Business Intelligence.

The need of the hour is to have a more people-centric, personalized view of analytics that doesn’t blindly believe in data, but to blend in people’s confirmation biases (based on social context – We find what we generally search for and We like to stand confirmed with the data that we find) and big picture (The same data could translate to a different interpretation in a different context) in which data is analyzed and interpreted.

Another news that caught my attention this week on analytics was around disaster modeling. A company named - Eqecat - does catastrophe risk modeling that helps to forecast losses from catastrophic events such as Hurricane sandy. This forecast helps government agencies and insurance companies for further planning and action.