Our organization has completed our initial ontology and taxonomy diagram for our enterprise. Internal discussions have come up with two (2) different flavors of what the raw vault should look like. I’ve attached both diagrams to this topic to get viewpoints from this organization to prevent us from going down a rabbit hole.
This diagram was believed to be too complicated to support because of the hierarchical structure. It went against our objective of limiting the customization of ingestion packages and would not scale even though it satisfied the relationship desired for retrieving the data at rest.
This image is what the majority believed embraced the standards of the data vault methodology. Those who felt it wasn’t sufficient to represent the data at rest objected. It was pointed out that our objective is currently the building and loading of the raw vault and not the reporting of the data. The reporting of the data is addressed in the business vault and information marts.
Please review and provide your thoughts. For the group looking to have the data at rest organized for reporting, would there needs be address with bridge and pit tables.
Thanks for your feedback.
My forst impression of both option1 and option2 is that they are heavily over engineered. Are the hubs really unique lists of business keys? Maybe you can provide us with some test data that is going to be inserted into the model? It would be easier to argue with real examples.
Hope this helps.
Thank you for your reply. Your point is well taken but as I stated in the first paragraph, this is our INITIAL approach. We haven’t progressed far enough to get into the data domains other than identifying the nouns we’re dealing with. My post’s objective is to confirm that our understanding is on point.
hmmm i wonder — would you ever categorize streetnames, numbers and suffixes as business objects? They are not — looks like an Ensemble Data Vault model.
Awesome DV icons btw
They’re not business objects and yes as AHenning stated the diagrams are over-engineered.
We were tasked with modeling the nouns to get to a 3NF diagram. The emphasis behind this was to satisfy the people in data governance. In any event, our next step is reviewing all of the diagrams and creating a business object physical DV ERD.
Oh yes – the DV icons are awesome.
I second (third) the question about whether Street Number and Street Suffix are in fact discrete business entities. What questions would be answered by having such a list in the Data Vault? “Generate a list of all addresses in London where the house number is 123.” “What are all of the addresses for houses and buildings where the address is on a Lane?”
Begs the question of how those Hubs would be useful.