On big data, a joke recently broke: at an internal industry conference in the movie industry, a giant movie industry spokesman said: through data mining, we found out the relevant selling preferences of different audiences. For example, the audience of Fame consumed more hot drinks than the audience of Wolf. These are things we didn't know before and couldn't have predicted.

A conclusion like the one above, which is based on an analysis of the viewing data of two films, seems to be objectively correct, but in fact, it is a joke because of imperfect modeling (lack of consideration of the viewing season) and other reasons.

In the recent past, when we took stock of FinTech, we found that big data itself was an "embarrassment". We searched the news and didn't find anything particularly worthwhile to say about the term. Only a little current affairs information to cobble together this keyword inventory.

In 2017, big data was so important, yet so uninformative.

Big data model is not perfect, because the roots are not firm big data has been lukewarm, and his development flaws have a lot to do with it. Although people are extremely optimistic about it, but failed to usher in the explosion of the industry.

Chatting with some friends who do big data, they will even be very blunt to spit out their own home data model.

"Those so-called data models and all that crap, you just need to glance at them and get a headache all day. The data in the model is huge and the clue logic is complex. A lot of data seems to be very important but extremely boring, meaningless to the judgment of the results, tasteless to eat and discarded, the existence of chicken ribs in general."

"Frankly speaking, the fundamental reason is not the backwardness of the technology, but the development of the whole industry's roots are too shallow to errata, generalize and rationally explain the validity of the data."

"Roughly speaking, a reasonable big data architecture is that the data model is perfect and can make a comprehensive and reasonable data streamlining according to the specific field, remove irrelevant and interfering data, sort out a reasonable and objective recommendation, and according to the subjective judgment and errata of the data analysts, and then summarize the reasonable conclusions, and make an accurate prognosis for the relevant industry. "

"Now what? Originally the data models were flawed in one way and another, but still thinking about the complete automation of data processing."

"And relying entirely on objective data to accomplish so-called AI algorithms is all bullshit."

"That joke about 'Fame' and 'Wolf' just now is actually a seemingly objective, but actually ridiculous, conclusion of the analysis."

"That's because, when people talk about big data, they take the data too much for granted. If you only rely on this awareness to do data analysis in the field of consumer finance, there must be a lot of investors are pitched to the bottom!"

"So now earning money or those who rely on the data companies that buy and sell user information, a data package, add some moisture, sell everywhere, unlimited earnings."

"However, recently it seems that it is not so easy to straighten out, because the official investigation is more and more strict, some of the so-called big data companies can not engage in, I am afraid that it is cool."

The Internet of Things may be a real opportunity for big data companies "In addition to the accumulation of industry experience, there is a need for more data for online support."

"Of course, it does not mean that the more data the better, but rather, the richer the data online, the more conducive to our organization of effective data."

"The core problem lies in how to generate large amounts of effective data."

"Effective data, simply put, on a certain field, for example, consumer finance field of a small segment of consumer goods related data, after a reasonable combination and deconstruction, to make a reasonable prediction of the industry's development, responsible for the expectations of investors data. Otherwise, the bigger the data, the heavier the burden, the less it becomes."

Accumulation of experience to when it is a head?

"Maybe not until the real age of the Internet of Things."

Why?

"The Internet of Things will allow more consumer finance data and logistics data to be online, and personal consumer credit information will be further online, and data aggregation and processing will be more efficient and comprehensive."

"However, with the rapid development of mobile payment, more people's financial consumption ability online is basically presented, including personal consumption habits and personal credit information are online, and the resulting logistics information, housing, loan information, etc. are gradually completed the ultimate online, these are excellent opportunities for big data. "

"The big data industry has great opportunities, but big data is an unstable industry, because all the data are attributed to the machine, and the machine is controlled by people, and the associated operational risks are entirely dependent on one's own sense of risk and character. The industry is ready to erupt at any time large-scale risk, good luck only affects the security of data, bad luck, very business and personal credit will go bankrupt. This will bring a huge disaster to the industry, and even society as a whole."

"Therefore, the relevant guidelines for practicing enterprises need to be further refined and regulated, and there needs to be a professional conduct aspect of regulation for people."

What kind of people use data how, for different purposes and to different effect.

This is again somewhat related to a big data-related paragraph, just the beginning of the paragraph, the end of the joke, and also quite complete.