The business drivers that we discussed earlier lead to the corresponding use cases. In some situations, business drivers and use cases are similar.
Metadata management (MM)
MM assists in performing the following business tasks:
- Document and manage organizational knowledge of data-related business terminology
The challenge to “speak the same language” has been on many companies’ agendas for years. Is it simple to create a common business glossary for the whole company? Unfortunately, no. An organization consists of multiple business lines and functions and often has its own business languages. The idea of creating a common enterprise data model has been replaced by the idea of multiple data models linked to each other. This concept is the core of the domain-driven design approach and data mesh architecture. So, what is the role of metadata management in this respect? If you recall, data models are metadata artifacts. They use business and technical metadata to describe, classify, and design data.
- Collect and integrate data from different sources.
An organization uses multiple internal and external sources. One of the biggest challenges is data integration. It would help if you understood semantics to match compatible data. You also need to know data formats at the physical level to implement integration tools. Data integration seems to be a “mission impossible without knowing the business and technical metadata.”
- Establish and enforce the use of technical metadata standards to enable data exchange.
Data should be exchanged with various stakeholders within and outside an organization. The organization should create technical metadata standards to implement the exchange.
- Ensure metadata quality
Metadata management should plan, implement, and maintain metadata in various repositories. The success of any data and metadata initiative depends on the quality of metadata. Data quality is one of these initiatives. Metadata describes data. Data quality requirements are examples of metadata. So, to manage the quality of data properly, data should be correctly described and modeled.
- Provide standard ways to make metadata accessible to metadata consumers.
Metadata sharing is another example of metadata use cases. All data users within an organization require different types of metadata to perform their daily tasks. For example, data management professionals need to know data models that represent business and technical metadata. IT professionals need to have access to operational metadata that demonstrates the performance of data-related processes.
Knowledge graph (KG)
There are two different types of KG, actioning and decisioning.
Actioning knowledge graphs are used to drive decisions or actions based on data. Knowledge graphs should link data elements from different data sources and make their integration transparent to enable decision-making. They have various business cases, including:
- Data lineage
Using knowledge graphs to demonstrate data lineage leads to a straightforward conclusion: data lineage and actioning knowledge graphs are the same thing. In this context, they can be considered synonymous.
- Impact and root-cause analysis
Data lineage and knowledge graphs allow for performing impact and root-cause analysis. Root-cause analysis assists in analyzing sources of data elements backward in data chains. If, for example, you need to investigate data quality issues or explain the amounts in financial reports, you use root-cause analysis. Impact analysis allows us to examine the usage of one element in a source system along data chains. Impact analysis is required in case of changes in an application landscape.
- Single view on a subject
Nowadays, many companies focus on getting the “Customer-360” view. They believe that it will improve customer experience and lead to revenue generation.
- Information search
For example, knowledge graphs can assist in searching documents and computing documents’ similarities.
Decisioning knowledge graphs assist in finding data trends. Examples of decisioning knowledge graphs are product recommendations, improving customer experience, forecasting business needs, and so on.
Data lineage
We have already discussed the key use cases for data lineage in the previous paragraph dedicated to knowledge graphs. These use cases are:
- Performing root-cause and impact analysis for multiple data management initiatives
- Explaining data origin and data transformation
- Integrating metadata
Conclusions:
- MM, KG, and DL have similar use cases: they all assist in collecting and integrating data and metadata.
- MM and KG intersect in documenting and managing organizational knowledge by adding semantics to physical data and metadata. In this case, KG can’t perform this task without having a proper MM in place: these two capabilities are two sides of the same coin.
- DL assists companies in performing impact and root-cause analysis and explaining data transformation. KG is one of the technologies that can be used to document DL. So, in this case, KG and DL are again two sides of the same coin.
- For more insights, visit the Data Crossroads Academy site://academy.datacrossroads.nl/courses/data-lineage-what-why-how/lesson/data-lineage-what-why-how/