Big data "too much to read"

According to the American http://www.aliyun.com/zixun/aggregation/31646.html > Media report, in the "Big Data Age", everything in our lives seems to be explained in a quantitative way, But as we become familiar with the term, we must also be wary of "excessive interpretation" of the big data on the risks to the business.

In the movie "Pat's Happy Script," Robert? The role played by Niro gives us a vivid "interpretation" of this. This is called the old school special? Solitano's role is a big fan of the Philadelphia Eagle football team. He believes that many unrelated, random events, such as the location of remote controls, can be a key factor in determining the outcome of a team's game.

Most people would think that the role of Niro was nothing more than a superstitious fellow. But in fact, his approach is in the eyes of others, it is too much interpretation of things, far-fetched so that the abuse of the performance of a random event that has nothing to do with each other to "unscramble", hoping to find out their relationship, and find out the "model" approach. And like to interpret seemingly meaningless things, it is the inherent characteristics of the human brain, I believe that many people have had this experience, by the subconscious influence, we always think a cloud shape is like a human face, always feel a piece of music behind the hidden information.

However, the researchers believe that this "excessive interpretation" of the behavior, is likely to cause harm to enterprises.

"The advent of large data has allowed some researchers to think that they can analyze anything by a certain amount of data." "Microsoft Research lead researcher Danah Boyd, New South Wales, State Of University associate professor Kate Crawford wrote in their published article," Because of the huge amount of data accumulation, there may be a link between the things that have nothing to do, which triggers the big data ' The tendency to read too much. ”

And enterprises are affected by this trend, will go to use large data analysis, to come up with some actually inaccurate conclusions, this analysis process itself not only cost, the conclusion will also affect the enterprise from advertising to management of many decision-making links.

A typical example of this is Google. Google developed a "flu trend" in 2008, a tool to determine the geographic location of the flu population by using Google's search for words such as flu, signs of infection, diagnosis, and so on, to arrive at a real-time map of the flu outbreak. The tool has been good for a while, but it went awry in December 2012. According to an article published in the journal Nature, Google provides twice times more flu-patient statistics than the Centers for Disease Control and Prevention. Why is that? The researchers believe that because of the "overwhelming" coverage of influenza outbreaks in the media, people's search for flu-related content has increased significantly, which makes Google's data excessively "inflated". If a pharmacy uses Google's data to determine the amount of inventory of flu-controlled drugs, it is likely to lead to excessive storage.

"More and more business brands are starting to rely on data to manage their relationships with their customers and to drive their business forward," he said. In this case, it is really daunting, frankly, to rely blindly on data to make decisions and to implement such decisions. Colin Fervent, general manager of the UK Market Research institute GfK NOP, said, "If companies stop and think about how these decisions are made, they are likely to suffer commercial losses." ”

"The use of large data is still subjective, and the quantified results do not sometimes represent objective facts." This is particularly true of social networking information. Danah Boyd and Kate Crawford said.

It is reported that, although people are gradually starting to have a cautious use of large data awareness, but for a reasonable use of large data skills, the lack of talent, but it is possible to "over read" large data situation intensified.

According to a 2011 report by the McKinsey Institute, the demand for "big data"-related talent with analytical skills such as quantitative analysis will reach 4 million by 2018. In addition, there is a risk of a 1.5 million employment shortfall in large data-related management and analysis personnel.

"At present, the most common complaint from customers is that they cannot find the right candidate." "The head of an agency responsible for data-related job openings told reporters.

According to a survey conducted by the agency, 89% of people in the data analysis industry typically get new offers every month through sites such as LinkedIn, and 25% of people are contacted by other companies as much as once a week. The person in charge is also predicting that with the development of nanotechnology, the demand for relevant jobs will be "blowout" in the future.