So I am talking with the team and this is not something that would be straightforward for so many reasons. We also moved it to Contribute > Feature because that shows more of what this would be.
Currently, those of us that work on/with data inhouse mostly use the models that are available in the source code:
I also got a look at the Northwind data model:
That is definitely easy to understand and fit on paper or one screen. 13 tables in all.
Comparing that to Discourse, we have a whole lot more tables over 180+ or more, visualising this would be … a journey. Especially because there are also tables from plugins (and these change from install to install) and data in *_custom_fields tables which should be included as well if you really want to have a complete picture.
Also because of how our database is designed, we can’t use most tools for data modeling, we’d need to find one that works with ActiveRecord models. And I think that also makes this tricky too, all these data conversations are over my head. 
But that is not to say this isn’t something we don’t want to do though, this is just commentary. I’d love to hear suggestions from you or anyone else on ways we could make this better. 
