Big data governance is not for everyone
The author combines the experience of data governance in the field of regional medical care and takes Kimball data warehouse construction methodology as the core. He hopes to explain the ideas and methods of big data governance in a simple and simple way, so that everyone can have a comprehensive understanding of the function and utility of big data governance, and get some ideas. Misunderstanding. (Kuo Haixing Shen@CSDN original, please indicate the source for reprinting)
1 Is it really difficult
1.1 What is Big Data Governace
Big data governance is different from business systems. It is not application software itself, without specific scenarios, pages, and functions. It is an excavation of the value of data in existing business systems. Therefore, the premise of big data governance is to have data, and the amount of data is large, the types or forms of data are diversified, and the content of data is constantly changing. The essence of data governance is the process of monetizing the messy data that originally existed in various business systems.
1.2 Big Data VS. Data WareHouse
Although, the concept of data warehouse predates big data, and it is not necessary to use data warehouse to mine the value of data. However, with the rise of cloud platforms, it is a general trend to use the ideas and technologies of data warehouses to solve big data governance issues based on cloud computing. Below, I will give an example.
In the scenario of regional medical data governance, the sources of data include: municipal hospital business data, primary medical public health data, mother and child health system, population death system, smart life channel system, operation supervision system, etc. The amount of data is very large, data sources and types are extremely numerous, and they are constantly changing.
After these data are managed, the application systems that need to be supported are: patient index, health records, prescription reviews, DRG, medical record quality inspection, etc. Will produce great value.
- standardization
Analogous to the process of transforming iron into iron, the same person’s inspection data (such as blood pressure measurement) may exist in the information systems of different manufacturers in various hospitals, facing various problems such as different table structures, different data units, and different code tables. Need for unified standardization;
- Modeling
It is analogous to the process of rolling steel into different types of steel billets. Although the data has been standardized, the data with the same meaning still exists in different tables, and the data is still messy. In order to facilitate data access, a model covering all data is designed according to business characteristics, and standardized data is injected into the model to form a data detail layer. At this point, the disorganized data has become unified, complete, standardized, and model data that can be easily accessed.
- Theming
It is analogous to forging steel blanks into standard parts required by different industries. If the business scenario is very complex, the modeled data needs to be themed according to different application scenarios. In the process of data theming, although there will be redundancy, because it is developed for specific application scenarios, the data will be like standard parts and can be used directly.
- Functional
It is analogous to the final production of a steel work, using modelized or themed data to develop data applications, and finally generate a data mart to form value.
The process of using data warehouse technology for data governance can be analogous to the manufacturing process of steel products. Through standard processes, it is convenient for data access, which not only improves development efficiency, but also reduces double calculations and saves calculation costs.
Seeing this, the logic of big data governance is clear, the technology is mature, and it doesn't seem to be too esoteric. But is it really just that?
To Be Continued…