This is my eighty-fifth original article \
Choosing a name is a science, and it’s really hard for science students to figure it out. Like this data map, lineage analysis and data assets. If you’re not in the data business, you wouldn’t think these three words were related!
The data map
Data map is one of the important functions of data governance. Is it, as the name suggests, a map of data? Yes! It’s a map of all the data that’s been managed. This map mainly solves the following problems: \
1. How many data resources does the platform have?
2. How many tables, fields, etc. does each data source have?
3. What is happening in these tables and fields?
How can I find this data?
5. What do I make of the data?
In general, data maps should have the following functions to answer the above questions:
1. Data overview
2. View metadata
3. Data preview
4. Data directory
5. Data retrieval
6. Data annotation (metadata management)
7. Blood relationship analysis
The names may be different, but the problem is the same. The idea is to tell you what data the platform is managing right now. Similar to the company’s inventory, ledger.
This is ali data map overview page, very fuzzy, see a general meaning on the line.
Blood analysis
Blood analysis, also called pedigree analysis and blood relationship, is one of the important functions of data governance and is generally placed under the data map/data management module.
Again, as the name implies, that is the data of the son, father, grandfather of the blood analysis? Right! Is the upstream and downstream of the data for context analysis! Consanguinity analysis mainly answers the following questions: \
1. Where is the data source?
2. Which table is upstream of this data? Which field?
3. Which table does this data flow downstream to? Which field?
4. What are the upstream and downstream tasks of this data? What task dependencies are there?
5. What are the possible impacts of this data modification?
All consanguinity analyses generally have the following functions:
1. Tracking of data sources;
2. Data impact analysis;
3. Task dependency analysis;
4. Impact analysis of statements;
Again, the names may be different, but the problem needs to be solved. Its core is to know the upstream and downstream relationship of this data/task/report, and the impact of changing this data structure/task/indicator/report on the upstream and downstream. Similar to the flow chart of a company, you can see which roles are responsible for the situation up and down, and which roles/people will be affected once the adjustment is made. Again, here’s the Ali’s Blood Ties feature page:
Data assets
It’s probably the word you see the most. We are also, as the name implies, a company’s “data” form of assets. This data asset mainly answers the following questions:
How much valuable data is available on the platform?
2. What is this valuable data available?
3. In what ways can the available valuable data be used?
4. How to control the permissions of these available valuable data?
So data assets generally have the following functions:
1. Data Asset Catalog
2. Data asset query and preview
3. Data asset permission application
4. Use of data assets (APIS, subscriptions, etc.)
A data asset is similar to a data map, but it’s all about data. Data assets focus on how valuable data is used. One is production-oriented, one is result-oriented, one deals with upstream and downstream dependencies and impacts, and one deals with how data generates value. I likened the data map above to a company’s inventory/ledger, where data assets are a list of currently available items. \
Of course, now some data center products in order to support the huge data asset management, will also make a data asset map, convenient resource search.
In addition, more and more data products now expand the concept of data assets to include data maps, which should be noted. This is also ali’s product, the function page of data assets:
If I have not made it clear, you can leave a message on the background of the official account, and we can talk privately
Enjoy better with the following articles
Data package | 】 OKR completely the ground practice handbook
“Dry” | traditional number what is the difference between the warehouse and the large number of data warehouse?
“Dry” | monthly salary of 3000 and 30000 data analyst is poor?
I need your upvotes. I love you
