Author | Shadow Attitude
What about big data is, from the beginning just to do data only, to begin realize that they have to do is central to the whole group of data products, and then went outside, the external output data capabilities, all process, will continue there are clients asked this question, there are ignorant, there are well-intentioned, but also provocative disdain. Like big data itself, but also with confusion psychedelic colors.
There are big data to understand the two extremes: stand outlet big data, many people feel that big data is the future, can solve all the problems, including artificial intelligence ; I do not think I would like for big data already decided everything good up; there is a, is the concept of big data speculation for several years, left - are copied into rice crust, but only extraordinary attempt to collect together the tens of thousands of points of damage, but so that large data
Big Data does not have to reach much data can be regarded as a large amount of body data, but if the data is less simple to just use xlsx, SPSS and other traditional data processing can be solved, then you do not need big data technology, little chickens with feeling chopper. But the application must not only refers to data analysis, out of the scope of traditional data analysis, large data applications can be used to guide all our work life business scenarios areas, such as personalized recommendations, precision marketing, risk monitoring and so on.
In recent years more or less settling down experience, and being humble when time and when the boss to challenge the summary of experience, plus the occasional cramp want to increase cultivation stumbling read the book, there are a few basic features about big data and we can discuss :
First, non-competitive
Now many companies, data protection is very strict, we all feel that their data is valuable, not just for others to use. There was a time, I was doing when data Union (This can be later opened with an article about), and each time the customer chat, wants to convey a concept, if the data is a resource, it can also be consumed by many parties at the same time use and re-use of resources, the use of the individual does not preclude the use of others. Only the data value is used to accumulate data (data directly selling naked of course is not the way big data application areas I have discussed).
Second, the value of no fixed value
Continuation of the above said, the bare data directly sell, after we have accumulated a great deal of practical experience in the application of data and then look back, I feel very funny, all want a certain pricing data field, and then publicly traded, they are short when and alienate behavior (personal point of view, do not move bricks). Data like intangible assets, the market valuation depends on the buyer's needs data, appropriate data services applied to a customer's product, produced vastly different role, insists on giving a data set a price, only allows high data devaluation or let buyers distrust of big data . But not to say that the data is priceless, the data must have a value , then let it flow applications up and constantly adjust to the appropriate value iteration, intelligent application's data will maximize the use of application mode has a real added value creation wealth.
Third, ask questions
Can not automatically obtain economic value from big data in a way, no shortcuts and take it for granted, I am against all that as long as there is a large data did not need anything to say. Simple data storage business not useful, but what to store, what is clear, the industry can choose the high-end technology are ready to wait for selection. Now for an enterprise: a prospective in-depth understanding of what data is first stored and processed worth, it is the top priority.
Big Data provided only answer, but it is the silence to ask the questions that come from the person in charge of intelligence, depending on what the key question! Like Plato put it: The ultimate winner will be those "who know how to ask questions," only those who know how to ask questions to know which data may answer the question of how to use these data to answer questions.
Big Data industry is most lacking or in the future the highest social status of people, is not only understand the business, but also understand the data, know how to ask questions of business, but also know big data experts how to answer questions with a big data (make ad: Welcome to count Lan candidates product manager upgrade worth)
Big Data technology provides fast high-end cutting-edge technology, data processing technology can make big traveling thousands of miles. But if you do not select the correct and effective direction, but the cumulative error for the cost paid. In the era of big data, a right direction, the right questions, the right ideas, massive real-time model algorithm is more important than sophisticated science and technology.
Fourth, the data must be active
Long story short, data shelf life, require constant updating, no data is her unchangeable, if the data is not updated, like stagnant water , so if there is data, it is necessary to use them quickly, keep the do not know how to use, keep for a long time on the value of the index declined.
Fifth, do not sell directly
Perhaps you think direct selling now form a chain, and that which is a new model, I just want to talk about the drawbacks of direct sale:
(1) the data itself is no exact pricing, pricing too early will reduce premium space;
(2) can not be sold by way of technical barriers (discussed today is the big data business applications, scientific research is another topic), data can be purchased cheap to sell again, creating a vicious cycle do low data value;
(3) data security issues, the individual data can not be applied by way of direct selling, but individual data is very worthy of study and data content applications, need to find a safe and lawful manner desensitization, the user-friendly big data to feed daily life .
Sixth, the correlation rather than causation
The traditional Chinese concept of cause and effect speak, so "big data correlation and not causality care," the idea came from a foreign country is still very subversive, but because of its powerful big data "extensive data can be calculated" feature, so that no need to explore cause and effect, you can make sufficient analysis or prediction based on the existing vast amounts of data. But this feature is not even on subversion once again, figuring out cause and effect, I still practice experience also welcome friends throwing bricks.
[Articles from a love of people along the way data essay]