? Every organization needs to perform an assessment of its data governance maturity, preferably annually.

Big Data modeling is a process of data mining, which is the process of discovering problems from the data, interpreting those problems, and modeling the data accordingly. Big data modeling is not just a technique but a process that is designed to solve the problems of business processes, and if there is no goal or if it is not aimed at solving the business side of the problem, then there is no big data modeling.

Big data modeling should be based on business understanding of business knowledge, knowing what is the relationship between this relevant data and the business problem and how it is related, and in the final stage of shaping, it is also about using business knowledge for modeling, and building up the big data model by asking and answering business questions.

Doing big data modeling, not just modeling this one action, the whole process of multiple links are very important, in the process of big data modeling, to find the right data source is the focus, for the data source for pre-processing is the difficulty, data pre-processing is difficult, although it is said that now there are already a lot of automated data processing tools can be used, but these analytical tools as well as a variety of analysis methods are also through a long period of exploration time.

When doing big data modeling, you can't rush in the data preprocessing stage, and you have to find the right analytical methods for data preprocessing. When doing big data modeling should focus on some of the original patterns of the data;

For example, in the process of customer purchase behavior analysis, can customers after the purchase prediction may be related to the previous purchase behavior, of course, this process and the operator's experience has a great relevance, especially in the understanding of the beginning of the business knowledge, may be a better understanding of this original pattern.

A model is built, many people will follow the model for a variety of predictions, if the prediction is accurate, it shows that the model is a good model, is valuable, in fact, this can not be used as a criterion for judging the value of a good big data model is in order to change the behavior of the business as well as the results of the prediction to improve the behavior of the business, the transfer of new knowledge and insights, as well as the need for the development of the business is it will be adapted to the needs of the development of the business. The need to adapt to the evolution of the business is its yardstick.