The New York Times as a long article on building the Spy Blogosphere - "Open Source Spying", to harness the interrelationships of information that exist within the intelligence agencies.
If you could get all the intelligence analysts to become bloggers in the Spy Blogosphere and then map the interrelationships of data you might come up with diagrams like these, below:
Open-Source Spying
Maybe that's what they have got now - I wonder if they can tell a threat by the geometric shapes that are created by the interrelationships of data?
In a way, those geometric relationships exist everywhere. For example, when artists drew and painted they'd use, consciously or unconsciously, mathematical constructs like the Golden Mean, or, as Cezanne's famous statement to see the world in terms of the cylinder, sphere and triangle.
Information also can be seen as a cylinder, square and triangle and that's what the sophisticated textual analysis mapping programs are trying to do, in my opinion. What they haven't yet decided - what the shapes mean in and of themselves.
It's one thing to figure out a relationship - it's quite another to decide if it harmful, beneficial or neutral.
If you get it wrong - you end up putting people in jail that are innocent or missing the people trying to do harm.
You'll notice the picture in the upper left is a dodecahedron
I think if the relationships can be mapped (including Web Analytics relationships - like Unique Visitors that go to a part of the site and do something - IE: a conversion metric) into geometric structures that have meaning - you can begin to take Web Analytics to another level.
Even something as disputed as Astrology had a belief that certain ratios were more stressful than other (IE: 90 degrees is more stressful than 60 degrees - in a relationship of two objects).
So now the Intelligence agencies are mapping relationships of information - it can't be long before we'll have that capacity in Web Analytics.
In fact, certain web analytics packages can chart data in geometric patterns - but I don't think they have decided if the relationships of data your filtering on are positive, negative or neutral and adjusted the geometries for that - once they can do that (as social networking software like Brand-Pulse does for filtering information in blog posts and message boards into categories) you'll get much more intelligent maps of data.
