Big data is definitely one of the hottest issues these days, but what is big data and what makes everyone so excited about it?
Big data has been defined in many different ways, and the most commonly used description is the 3V (Volume, Variety, and Velocity) model.
This model defines the large volume, exceeding the limits of existing software, variety that includes not only structured data but also unstructured data with text, and high velocity which changes online in real time as three components of big data.
Big data consists of a large amount of various forms of information including both structured data and unstructured data (such as text and documents) from SNS and social media.
Some of the most popular methods are the data mining method that extracts necessary information from a massive amount of data, machine learning which is able to process online data newly added repeatedly, and morpheme analysis for unstructured data such as with online documents.
The current job of big data analysis combines machine learning and data mining methods to extract important information from an unrefined sea of data and finds patterns using a pattern recognition function.
Let’s suppose you’re at a market shopping for apples. It can be difficult figuring out which ones are good or bad. With big data analytics, you can sort out different types of apples using a pattern recognition method.
The Good: Customized Services
From the point of view of clients, “the good” would be the services that perfectly suit each and every one of them. The most well-known success stories of customized services are Amazon and Netflix.
These customized services using big data analytics can be provided offline as well.
For example, department stores only collect the total number of purchases on their sales accounting systems. They usually don’t deploy marketing strategies to categorize customers and sales patterns between various shops.
A Korean department store recently began broadening the information collected through CRM (Customer Relationship Management), and to categorize their big data for diverse items such as regional sales patterns, net profits, and frequency, instead of just viewing it as a single sales record.
They also intensified interoperability between business and information so that data can be brought together and effectively utilized on the system, while creating a strong cooperative system across different departments for better integration.
They created a big data system, analyzed purchase patterns of various types of customers, and used the information to make suggestions for related products or to induce secondary and tertiary purchases.
One of the patterns they found was that sales tend to go up at grocery and children’s departments whenever it snows. By creating a specific booth for groceries and products for children, mothers are able to locate these items much more easily, and the department store can increase its sales on snowy days.
The Bad: Risk Management
Big data analytics not only suggests the best services, but also finds what may cause harm and uses this information to warn and protect clients.
Google Flu Trends from the U.S. Centers for Disease Control, which figured out and warned regions that were highly vulnerable for the flu two weeks prior to its spread, is one of the most prominent examples of big data analysis.
The Korea National Health Insurance Corporation is running a similar disease forecast service as well. The basis of this service are national health insurance records, which consist of a massive amount of big data consisting of emergence, progress, and treatment information of diseases experienced by 50 million people.
The disease forecast model is developed through big data analytics which conflates the national health insurance information on epidemic diseases and data from social media. The service not only calculates the risk level of major diseases, but also provides regional and age-specific warnings for possible epidemic diseases.
The Ugly: Typology
Did you know that Obama’s camp used big data analytics for their 2012 presidential election campaign? They went a step beyond the old campaign methods and analyzed the voting list.
They first categorized the voters to people who can make donations, those who can volunteer to help campaign, and to those who may change their minds. This resulted in creating a different strategy for each group, and these strategies were designed to be altered according to repeated voting forecasting simulations.
As we all know, the winner of the election was the Obama camp, which established a strategy using big data analytics.
Just like the phrase, “history is a mirror to the future”, big data and big data analytics is a mirror that shows us the future and helps us prepare for it. As more data is being accumulated rapidly, we will be able to see the future more clearly.
Written by Jinsu Kim, LG CNS Big Data Business Unit