Sunday, June 28, 2015

Learning from the past

There is this saying about not learning from past, wait what happens to people who actually learn from past?





TAAA DAAAA! Yes they can predict the future. We all have heard words predictive analyctics, machine learning etc. What exactly are we talking about?

Wikipedia says : "Predictive analytics encompasses a variety of statistical techniques from modeling, machine learning, and data mining that analyze current and historical facts to make predictions about future, or otherwise unknown, events"

Yes. Now that we have tons of data , how can we find patterns and learn from the past. As George Santayana rightly said, the ones who do not learn from the history are ....

Crazy analytics comes into picture here. A lot of data modelling and research is going on and honestly I am not the best person to talk about it.

What I simply understand is its equally important to process the data in a right way and make sure sensible patterns are generated from the patterns. This is where the values lies. The unused data is as good as garbage.

Lets go through some use cases of predictive analytics 


Use case 1 : Network : Definitely my favorite one! So what kind of  data comes here ans whats the prediction we are talking about. So for network analysis the data is coming from network testing equipment , MME logs , all the call data records, data in the OSS. Now there is literally huge data if all of these are combined. We are talking of 100s of thousands of sites. A platform that will take the data from all the possible sources and analyse it. Over the period of time, some patterns will be seen from the data and definitely some predictions can be made. There is a lot of scope in this field and with SDN (Software defined network) the network will get the flexibility to implement such changes.



Use Case 2 : Baby monitoring systems : If the trackers track the time when your baby sleeps, drinks , yawns (why not!) , cries , plays. Over the period of time you will be able to know the moods, sleep cycles. And also the mothers will get to plan their activities accordingly. This data will be more and more accurate over the period of time. (I am not sure if crowd sourcing will help here, for example using the data from babies with the same age and their sleep cycles, I am not sure)


Use Case 3 : Connected House : Yes the highly useful one. What time one leaves to office, what is the optimum temperature set in the house as per the different weather conditions. What are the food habits as per weather. Who are frequent , known visitors, who are the banned ones! The grocery stock, automated promotions. (Guess when you are almost running out of detergent and you get a promotion from Tide, which brand are you going to buy???). Home organisation, closet organisation (Very difficult to understand the pain unless you actually have atleast 10 shades of black dresses in your closet! ) , Food choices , family time monitoring. Automated lights and the list will never end,