How to build an enterprise-level data governance system?

Editor’s Guide: Data management is a full process of construction and management of corporate data architecture, data standards, data quality, data security, etc. As a data worker, whether the company’s data volume is large or small, it should have the ability to manage data. From our own experience, this paper analyzes how to build an enterprise-level data governance system, hoping to help you.

Usually, when I am doing a data architecture, I will add a “data governance” next to it. But it, most of the time is just writing. Because this data is too strong, it is not expensive in the short term, and it can take effect for a long time, but it is not enough for the outside. Just like a good wife, a clean, net Lisso, who is picking up at home. Older seems to praise a “true clean”, but it will not say that you are big.

The above figure is a data architecture example, and data management is a very important part. Its meaning determines the availability of the upper application, even determines the success or failure of the project. After many data, the boss has a very significant feeling: there is no secretary to spend hundreds of thousands of systems. Is this awkward?

1. What is the use of “Data Governance”?

An explanation: “Data Governance” is big, it is a full process of the construction and management of the data architecture, data standards, data quality, and data security of the company; Plan, supervise and control.

Well, you understand the right, usually the company’s company, data governance will be put on the desktop. But as a data worker, I think there is no matter how much data do you have – even just a remembering – you need data management. A pursuit of perfect data workers should have data cleansing.

Second, what should “Data Governance?

In fact, the vast majority of things have long since study, and there are many research results. There will be a lot of information for a slight collection.

At present, the theory and norms of domestic and foreign data governance:





In the data management capability maturity assessment model, the data management capability is defined in the data management capability, data standards, data management, data application, data quality, data security, etc. Eight aspects were evaluated.

In the Dama-DMBOK Function Function, 10 major data management functions are defined:

These two figures can be posted to your PPT, let your leadership look at how data governance is a huge system.

Third, what should “Data Governance?

Snoving you telling you a tip, there is a big hat cover in front, it is convenient to do things, this trick is not trying.

Two pictures in front, give you a huge hat, but how to implement it? This can be large and controlled.

I will give you another picture:

If you don’t have a few data engineers if you have a few data engineers. The boss said that you have to build a data middle desk, then you have two options: departure; persistence, must insist, insist on watching the failure of the project, so you can experience a complete failure case.

Pull back. If the company wants you to do data governance, data management, data strategy and other things, you will write it on the table above.

The first stage: actually is a variety of plans

The primary thing is definitely an organization guarantee. If you don’t say it, you will recruit a few people. The more people, your location is steady (from the Internet to touch the Dafa, laughing ^ _ ^); then a variety of processes, Standard, institutional construction, data security management system, data processing process, data extraction process, metadata standard, main data standard, etc. There are also a variety of touch, data resource carding, data demand carding, and more.

The second stage: actually is a variety of construction

According to the metadata standard, the main data standard, the construction metadata, the main data is based on the data processing process, the data extraction process, the data demand card results, the construction of various fixed reports, ¼´, inquiry, etc. according to the data resource card results, the number of markings And build

The third stage: actually is a variety of platforms

Construction of data floor map, blood analysis, data asset catalog, etc. Plus some unified controlled services, the data in the data is almost symmetrical, ready

Fourth, can data governance can you get in place?

See which step starts from.

If your company has a construction data team, it has already lapled awareness, unified definition, process specifications, rigorous system, rigorous technology, and guarantee, there is no problem, just organize the study of relevant knowledge, and be a half-year plan , Invested humanity, and certainly one step in place.

If your company is 3,5 data engineers, then you have two choices, leave; insist on watching the failure of a complete project. By the way, it seems that it is said.

Do you ask me to be so definite? Because I have experienced ~~

The foundation is unstable, no different in the mud pond, can be a miracle!

V. Data management panorama

Ok, finally put a data governance panorama, town building!

Forward, praise, look, arrange it?Author: big data architect, Director of China, China Director, good at Bi, Digital, Data Mid Taiwan Product Planning Field, Public number: Big Data Architect