Big Data Is So Hot: Advise You Not To Blindly Believe In Data.

Big data fever: advise you not to blindly believe in data.

Data are often misleading from specific backgrounds.

We must be very cautious about what inferences can be reasonably drawn from the data.

The mass killing of rats in Hanoi in 1902 is a classic reminder that we need to be vigilant against the data we measure and the incentives we offer.

At that time, the French colonial rulers were shocked by the flooding of rats in the sewer of the city, offering a reward to every local rat catcher every time he killed a rat.

Every time a rat tail is handed over as a deratization evidence, the municipal government will pay a penny.

At first, the data looked promising, but unfortunately, the plan went wrong.

The crafty Vietnamese entrepreneurs simply cut down the tail of the live mouse and set up a rat farm to increase their income.

A few years later, a plague broke out in Hanoi.

Today, the data generated by all of us on smartphones seems to be far from the statistical data of colonial rats in Hanoi.

However, there is no difference in the risk of misinterpretation of the data we produce.

Relevance is sometimes deceptive.

Incentives are always being teased.

Stripping the background, data may be, and often misleading.

Today, Alibaba, China's technology giant Alibaba, seems to be learning some of its lessons in testing its SesameCredit scoring system.

By collecting hundreds of millions of users' data, Alibaba has hoped to establish a reliable standard of consumer confidence.

Sesame is based on all the information from online shopping records to the cost of the subway, and then can be used to issue or refuse to sell - consumer loans.

But as the Financial Times reported, there is a significant difference between big data and strong data. Alibaba has not yet made use of its sesame credit score to issue loans.

As Dai Xin, a law professor at Ocean University of China (OceanUniversity ofChina), told the financial times, it is difficult to create models that have reliable predictive power under different situations.

Will "plagiarized students cheat too? Will companies that fail to repay debts go back on their construction contracts?" he asked.

A senior American Technology Corp executive who I spoke with recently explained that the algorithm was designed to discriminate and divide people into different categories.

But this means that we must be very careful in understanding which data is included and excluded in any given model and what inferences can be reasonably drawn.

Otherwise, in his words, algorithm discrimination may become "carbon monoxide" of big data - colorless, tasteless, but potentially fatal.

They become safe only when data are properly oxidized.

Steffen Mau, in his upcoming new book, The MetricSociety, outlines the greater risks of quantifying our lives in so many ways, and gives us an ominous warning. Steffen.

Our obsession with measuring everything - from academic achievements, personal appearance, behavior habits to popularity - is creating a new social order of values, a "follow the rules" culture, a "credible fiction" world.

Statistical data reflect not only the existing world, but also another new reality.

Data are used not only to provide information to society, but also to shape society.

HumboldtUniversity, a professor of macrosociology at Humboldt University Berlin, believes that this obsession with quantitative assessment can lead to material inequality being replaced by digital inequality.

The clash between classes will be replaced by competition among individuals - think of Uber drivers competing for higher scores.

"Number describes, creates and replaces status," he wrote. "Numbers make people."

Therefore, who decides which figures to collect and who decides the importance of these figures becomes a way of exercising power.

However, the methods used by organizations (whether international agencies, government departments, or global technology companies) to make such decisions are not subject to too much scrutiny (if any).

This becomes more and more important when the algorithm decides more and more what kind of achievement the student obtains in school, what kind of work the applicant can get, and whether the prisoner is released on parole.

One solution is to strive to subvert the concept of tracking technology and encourage individuals to create their own data stories to monitor and challenge the powers of those in power.

This may lead to a culture of anti monitoring rather than monitoring and ignoring (rather than monitoring).

How the data driven global environment movement has changed the debate on climate change is an encouraging example.

Or, some agencies may stop playing quantitative games, just as Gent University (GhentUniversity) seems determined to do.

In December, the Belgian university announced that it would dilute competitive and bureaucratic ways of publishing and citing indicators for making funding decisions.

On the contrary, RikVan de Walle, the principal, announced that the school will promote a culture that scholars have developed to enhance cooperation between research teams and faculty and staff members. Vander Valle,