Do We Want a Knowledge High quality Career?
In a current article on Knowledge High quality, Arkady Maydanchik urged {that a} key answer to the present information high quality points, that plague enterprise in every single place, can be the creation of a devoted Knowledge High quality (DQ) career.
I urged, in a reply to the article that, removed from fixing the issues, the creation of a DQ career was the one factor that may assure the perpetuation of knowledge high quality points.
Root Trigger Evaluation
For my part, the one acceptable argument for establishing a DQ career can be to remove all information high quality points, i.e. for the career to make itself redundant. Have you learnt of any current career, or are you able to even think about one, who can be dedicated to this? The sceptical amongst us would possibly recommend {that a} DQ career would possibly actually have a vested curiosity in perpetuating the issues. What are your opinions?
Most present Data Quality for Azure Data Lake points have their root within the separation in IT, for some loopy purpose, of the roles of these modelling operate and people modelling information.
Knowledge solely exists in an enterprise to assist operate. No different purpose.
Due to this fact, if you happen to handle the standard of the info created and reworked by operate, you handle the standard of all information within the enterprise. The present separation of the roles of operate and information analyst prevents this taking place.
Perform and Knowledge Evaluation Must be Mixed.
Each good enterprise analyst (and who would make use of a foul one?) ought to have the ability to do top quality information evaluation and each good information analyst ought to have the ability to do top quality operate evaluation.
And what would we name this mixed position? Enterprise Analyst. As a result of an excellent enterprise analyst wants to totally perceive – and which means analyse and mannequin – each operate and information. People who can not are solely understanding and modelling half the image.