current position:Home>What are the data quality assurance principles?

What are the data quality assurance principles?

2022-02-04 17:30:19 Alibaba cloud Q & A

What are the data quality assurance principles ?

Take the answer 1:

integrity .

Integrity refers to whether data records and information are complete , Is there a missing situation . Data loss mainly includes the loss of records and the loss of information of a field in the records , Both will cause inaccurate statistical results , so to speak , Integrity is the most basic guarantee of data quality . For example, there is a relatively stable amount of business data every day 100 Ten thousand records , A sudden drop one day 1 Ten thousand , Then maybe the record is missing . For a field in the record, the information is missing , For example, the score of an examination paper in the score table of a college entrance examination should correspond to an admission card number , The number of null values in this field should be 0, Once greater than 0, This means that the information is missing .

accuracy .

Accuracy refers to whether the information and data recorded in the data are accurate , Whether there is abnormal or wrong information . For example, the score in the report card is negative , For example, there is no buyer information in the order , These are all problematic . Ensuring the accuracy of records is also an essential principle of data quality .

Uniformity .

Consistency is generally reflected in data warehouses with a large span . For example, there are many business branches in the company , Consistency must be ensured for the same data . For example, users ID, From online business library processing to data warehouse , Then go to each data application node , It must all be of the same type 、 Keep the same length . So in 《MaxCompute Code guide for data warehouse construction 》 There is “ Public level ” Processing of , Ensure data consistency .

timeliness .

Ensure the timely output of data , Reflect the value of data . For example, decision-making analysts generally hope to see the data of the previous day on the same day, rather than waiting three or five days to see a certain data analysis result , Otherwise, the value of data timeliness will be lost , Make data analysis meaningless .

copyright notice
author[Alibaba cloud Q & A],Please bring the original link to reprint, thank you.

Random recommended