> It's just a jumble of poorly formatted text that isn't really contextually aware and is largely useless for the volume of textual documents.

I did a quick spot check and the lack of _clear_ date field is going to make contextualizing a bit trickier. It looks like most of the `email` have them but other types like `report` may have an unknown "first, created/circulated internally" date and a broader "the public can see it" date.

Nevertheless, it's only a matter of time before this gets loaded into a graph DB so the context becomes more apparent similar to what the journalists did for the panama papers.