Welcome to the spring of big data blossom

In the era of big data, the data is now a hot word, about having big data, how to play the value of big data, and talking about, and I think, it seems that this is a bit wrong cause and result, like the correlation, there is a when, b with it, and B, a but not necessarily associated, The author also describes what I think of big data thinking from a typical 4 v.

First, the amount of big data, the amount of data is large enough to achieve statistical significance, only value. I have seen a typical case is, for example, the traditional collection of thousands of data, it is difficult to find the effect of blood relationship on genetic disease, and once reached more than 20,000, then found that the effect will be very obvious. It is debatable, then, whether we are collecting problems in order to discover hidden knowledge to collect data or to collect it regardless of value. In fact, data collection, for the data itself, or can be divided into a number of standards, to establish a level, combined with demand, the goal to collect, of course, some people will say, this will lead to huge deviations, such as the loss of data integrity, there is a certain subjective bias, but I think, At a minimum, the value of the collected data can be relatively high.

The second is the type of big data, can also be said to be a dimension of the data, for an object, to take a tabbed way, to mark, for the needs of the kind of expansion, and the amount of data, I think the same is recommended according to the requirements to establish, but for the label, there is a commonly adopted strategy, That is the problem of recommending labels and custom labels, the taxonomy is actually a great innovation of human civilization, the way to adopt the recommended label, can greatly reduce the total amount of labels, and reduce the later protocol work, data collection, expansion, expansion of dimensions, but when the data into the application state, we want to deal with the small data, less dimensions , and through this recommendation, the alternative way, can be standardized based on the customization, rather than the non-rule extension, and even the user's custom tags to give some restrictions, so that the value of the dimension can be more apparent.

The third is about timeliness, now into the reading of the second era, then in a very short time to conduct problem analysis, association recommendations, decision-making, and so on, the amount of data needed and the type of data compared to the previous, often more, in other words, because the big data times time-sensitive requirements, so the way to deal with the data has changed, , now must become single-processing, single-processing, then the corresponding information systems, working methods, and even the organizational model of the enterprise, management performance needs to change, such as the author once worked on the enterprise, the ERP system, the designer opinion is very large, said a typical case, the past issued a change of single, sent to work to end, And on the ERP system, you must set up the material code for this change order, set up the need to query the storage of materials, and these are the previous designers regardless of, but not for the designers to pay for these increased work, and even because the lack of materials caused by the Change Order can not be issued, so that the designer work is not completed, resulting in punishment But since we've done it all at once, it's obviously necessary to improve the efficiency of your business, and the way that design changes are integrated with materials. Then as a staff, how to make their own work more comprehensive, more complete, to avoid the palace, so that the whole enterprise work more time-competitive, improve the number of data, type, processing capacity is necessary.

On the value of big data, one argument is that big data has great value, and one is that compared to previous structured data and small amounts of data, it is now big data, so the unit value of big Data falls. The author thinks both of these statements are correct, which is a question from the value of the unit data from the view of the overall value. And the author puts forward a new view on the value of big data, which is another way to really play the value of big data. This idea is aimed at the enterprise's problem, first of all to say what is the problem, the author said the problem is not the general sense of the problem, because a problem, we all think bad, wrong, etc., and the author's definition of the problem is the difference between the state and its desired state, including three models, the first is the usual meaning of the problem Must save immediately, in fact, this is the least one of the three modes; the second mode is to keep the state, and the third mode is the desired state, which is one level higher than the original state.

We propose a range of business intelligence product solutions for the problem, often in a variety of ways, such as employee training, such as device improvements, such as changes in how organizations are organized, and of course solutions that include information technology, big data means, and how we need to weigh big data is a relatively superior approach, If it is, then this means to solve, then it is valuable. For example, I know a case, an enterprise a product parts occasionally problems, the enterprise after several times decided to equipment on a set of industrial control system, recording material temperature, the results of another problem, the analysis that if the workers work normally, should not have such data records, And on the duty worker's question, the duty worker admits that he sleeps on the night shift, did not deal with in time. Again, the same problem never happened again.

Welcome to the spring of big data blossom