Democratizing Analytics –A Machine learning perspective

Democratizing Analytics –A Machine learning perspective

Hope my new born perspective of machine learning would be helpful as a bonus I have added a puzzle at the end of the article.

As we march towards 2020, large dataset are setting up traditional for new normal. The data is being generated by variety of sources other than people and servers, such as sensors embedded into phones and wearable devices, video surveillance cameras, MRI scanners, set-top boxes to name a few uff..

As we ride the cart wheel of digitization, data we generate annually - will reach 44 zettabytes, or 44 trillion gigabytes by the year 2020, which is ten times the size of the digital universe in 2013.

In one of the conference someone mentioned that Data is new gold mine. We fill lot of litmus survey, feedback, etc. in that case you are the product who is providing data. Apart from digitization of services and enterprises, a new trend has emerged recently to network all the man-made things around us, such as cars, home appliances, weapons, traffic lights, and power meters, yes you guess it right the new buzz world IOT.

However all the data we generate, are useful for descriptive or predictive analysis. Only a part of the data in the digital universe is useful when tagged, augmented. So what should I do with this Lake of data that that is coming in all forms Streams, Batch, Data at Rest and Data generated elsewhere?

With Machine Learning in the mix we can derive lot of insights from this like Descriptive (What happened), Diagnosed (Why it happened), Predictive (What could happen) and Prescriptive (What we should do) to understand thinks like Customer Behavior, Medical diagnosis and Map Genes to functions.

Let’s talk about Machine learning for a second and see how services like Azure Machine learning and , Python have democratize the Machine learning to derive insights. A quick scoop of machine learning does not harm like Supervised (I label the data and tell machine), Unsupervised (Clustering, ask machine to makes sense out of large dataset) and Deep Learning (pattern recognition, computer vision, natural language processing and speech recognition).

Next time when you get a call from you bank if you did a big transaction, or someone withdraws $500 before and after midnight, remember it’s not human its machine learning segmenting your transaction based on your profile.

Where is our way to future? Google, SIRI, IVR are already part of our mainstream, we are hitting our next step where sensors pick up our brain pulses and derive meaning out of it. Wearable for your brain like EEG from EMOTIV along with machine learning would make things happened based on our thinking patterns and needs.

The future is bright with Machine to learn from the data we generate and Predict and Prescribe. I think human is learning and trying to copy nature, remember neural networks in Machine learning is based out how our brain works….

With this I will leave you with you imagination what you can achieve with Machine Learning and Big data. Also to kick start you all with Machine learning, let kick start with a Puzzle.

 A coin is biased with Head probability (3/4) and Tails (1/4). I want to use this coin in a football game so that my outcomes are even, how may tosses do I required? Best of luck and Let me know J