Making sense of big data

3 Aug

Data on its own have only limited usefulness. It is only when we interpret them as information that we can understand what they actually mean. Consider a company knowing somebody’s date of birth. Unless we know also the date of this year and we are able to subtract those two to get the age of the customer, the data would be useless. Once translated into information (age), the company can tailor its proposition by e.g. offering a youth discount. That was easy, but how to make sense of huge amounts of seemingly unrelated data organizations are amassing these days?

The Economist published in May a very interesting article on big data and new businesses focusing on this market. Especially big banks are making use big data to detect fraudulent transactions. They collect lots of data about each transaction such as the amount, date, location or name of the sender. If everything was ok, the data would have certain patterns and would not deviate from them in the short term in a specific way. People would be depositing certain amounts of money in city A and sending e.g. 3% to city B in another country. Sometimes there are exceptions, but those happen on random basis. When analysing big data, companies therefore focus on deviations from standard patterns and focus on new emerging ones. If there are suddenly lot of transactions from the one city to another one, it raises a red flag to be investigated.

It’s all about patterns

I was reading this article when I was still living in China and thought how else can this concept be used. And I came up with one example. I lived in a city called Suzhou and every Friday evening took a train ride to Shanghai and returned back late on Sunday. When buying tickets I had to show my passport and they noted its number. If I was working in China on an incorrect visa, I would be in trouble. As an application of big data analysis, the Chinese authorities can for example track passport numbers of train passengers and look for patterns where there should be none. One would not expect somebody with a tourist visa to be making regular trips over two months – such person should be travelling randomly across the country. Unsure of how sophisticated they are I started using my driving license which looks similar to Chinese id card instead of my passport every now and then, just to be on the safe side.

Can you think of some other situations where you can find patters and how this information can be used?

 

2 Responses to “Making sense of big data”

  1. Patricio Zambrano August 17, 2012 at 9:04 am #

    Yo Daniel!

    Actually I believe what you are describing sounds alot to Marketing, us marketeers (currently doing a masters in such area) love to anaylze this behavior.

    For example, we implemented google analytics into our website (hotel) which allows us to see when people visit our site, how much time they spent on which page of the website and more. So the pattern I am trying to describe here is how the highest number of visits occur just after the 15th of each month. We realised that during the next 5 days or so we get the most inquires. I asked myself that question over and over again and later up I came up with a theory of the cause. Apparently most of the foreigners/chinese workers get their salary around the 15th of each month.

    So basically that is one of the patterns I have experienced. It is really interesting.

    Nowadays we are thinking on how to implement a newsletter with sale promotions to try and get this visitors or past visitors.

    Cheers from Shanghai bro!!

  2. danielvave August 17, 2012 at 10:00 am #

    That is a really good insight Pato! Thanks for sharing. I guess there are many more ways we can interpret data and use them to our advantage.

Leave a Reply to danielvave