My documentation style is pretty low tech as these things go. Nothing automated but it's easy enough.
I write most of my columns and measures out in Notepad++ and then copy it into Power BI. The advantages here are 1) you can format your code nicely without trouble and 2) your documentation is half-done already. Especially if you comment your code as you go.
Beyond that it's mostly a matter of organizing the code and documenting the queries. For the queries, open the advanced editor and copy all that code. You can later recreate any query instantly by pasting that back into a blank query*.
For the documentation itself, my preference is to organize by tables. A brief description of the data represented, the query, a list of the columns and their data types, then the custom columns in alphabetic order and the measures, same.
I tend to create master datasets and use them each for several different reports, so I tend not to document the reports themselves very extensively. I just keep a changelog of requests that I get about them, because everyone around here likes to change their minds about what charts they want to see. One changelog per report, starting with a note about which dataset it's coming from.
I've had to recreate one lost report so far. The original took about a week, the recovery took about three hours, and to be fair I was watching a movie while I did it.
*Assuming in some cases that you've already established a connection and credentials with the data source from the machine you're working at.
Did I answer your question? Mark my post as a solution!Proud to be a Super User!
