The ultimate Fabric, Power BI, SQL, and AI community-led learning event. Save €200 with code FABCOMM.
Get registeredEnhance your career with this limited time 50% discount on Fabric and Power BI exams. Ends August 31st. Request your voucher.
its looking to me like column descriptions (eg for a column called cost , "Based on standard cost") are set in the model view and then hovering elsewhere reveals the description. From a reusability perspective how can we help to ensure we wont be re entering these descriptions (maybe synonyms too) every time the same column or at least the same column in the same table is referenced in a model that is slightly different from the last one we created and therefore in a new dataset?
Solved! Go to Solution.
You have to assume that in a reasonably sized company (more than a couple thousand employees) Power BI and Azure are not the only tools and platforms being used in that company. Smaller companies may be a "single shop" environment and may be easier to manage from a data stewardship perspective. Larger companies have enormous amounts of data subjects, way beyond your example of sales / product management / manufacturing. They also tend to have a lot more politics and fiscal chopstick battles.
thx lbendlin, unfortunately (as far as i can tell right now) advocating as you suggest would equate to advocating not to use pbi in its current state...unless the new data mart feature (which i havent tried/read yet) starts addressing these kinds of shortcomings. In my mind, a certified sales query want to look up product info. So too does a certified shop floor query. I cant help feeling like I'm missing something.
You have to assume that in a reasonably sized company (more than a couple thousand employees) Power BI and Azure are not the only tools and platforms being used in that company. Smaller companies may be a "single shop" environment and may be easier to manage from a data stewardship perspective. Larger companies have enormous amounts of data subjects, way beyond your example of sales / product management / manufacturing. They also tend to have a lot more politics and fiscal chopstick battles.
thx ibendlin, i was getting more at their use in pbi. I'm a big fan of governance and all that goes with it. The problem i'm having right now with certified data sets (i havent looked at the new pbi data mart feature yet) is that they would likely overlap table/column wise with other certified datasets in the real world. Which means the same table/column would need a description entered twice (or more) if i feel that i really want pbi to also (in addition to dictionary like products) know what the unified definition of that column is. From what I've seen, dataflows cant be used to address this challenge.
both certified datasets and dataflows are tools. It is up to you how to use these tools.
The very nature of Power BI is to encourage and empower the "citizen BI developer" which immediately and brutally flies in the face of the more rigid and structured data governance proess that is preferred by the IT organization. It's a battleground, and you need to choose your battles wisely. Best you can do is advocate for and demonstrate the benefits of a solid data governance framework to that user base (and to management).
On that (small) scale this is entirely up to you. On a grander scale (enterprise level) this touches on data stewardship, data governance and data domain questions - these have their own tools like Purview or Collibra etc. It also can manifest in recommended usage of "golden" (promoted/certified) datasets etc.
User | Count |
---|---|
36 | |
15 | |
12 | |
11 | |
9 |
User | Count |
---|---|
46 | |
44 | |
19 | |
18 | |
18 |