Machine Learning, A Look in the Past

Before the Big Data become popular, there were at the back of Web 1.0 the machine learning of the past which utilizes Market Basket Analysis. These are very dominant in advanced e-commerce stores and online shops. The Job sites also utilized these technology before, and how did they implement it? Cookies, not those in your kitchen jar, but those text files that remembers your preferences, your visited sites and the things that you’ve clicked on the internet.

And what was that? Machine Learning, a part of the task of a so-called Data Scientists of today. Facebook analyzes all of our likes, shares, streams today, Twitter can also do it, I have even tried to do sentiment analysis of tweets using python. Google with their intelligent algorithms, Yahoo the early adopter of Hadoop for HDFS (a Big Data System). A lot of other database management systems like SQL are there used widespread. In those days, MatLab is a mostly used software, SPSS, SAS, S-Plus, and now R. Nowadays there is Pig to simplify MapReduce, the language for Hadoop management.

But who are those that have benefit from data science in the past? Amazon, the online book store have utilized data science, data mining, data analysis in order to show you the most relevant product that you can buy, they are now an online store and have even adopted into Cloud Service Provider company. Their algorithms can help upsell and show you related items to what you have already bought.

The most successful in utilizing BIg Data and Data Science is Walmart, they know how much to display on store, they know how much to carry on their inventory and they even know when you will buy your next coffee beans, sugar and even the infant milk and cereals that you consume and buy on your scheduled shopping. The likes of forecasting sales, that is why Walmart grew because of this so called business intelligence, it is data science, they use algorithms, mathematical equations, operations research tools in order to manage and understand the consumer behavior.

So the realization of Data Scientists today are thing of the past, but now, a successful e-scientist must have the skills in diverse fields (multidisciplinary-skilled) like business / marketing, economics, mathematics, statistics, operations research, some IT skills, big data and creativity. Yes, creativity, without it there will be no spark of wisdom, and this is mostly part intuition, insight and looking the world/data at a different angle to predict, to deduce and to induce.