I agree that geographical location is informative only, but the registrar data is definitive. If I have an intrusion detection system for a voting system in the US, and you see repeated intrusion attempts from an IP that is registered in Russia, it is highly likely *not* a local registered voter. (swap any of the two countries with random other two countries)

You don't need an exact location for that.

And yes, that is one of the experiments I am doing. Nothing is automatic (yet), but reports are quite informative this way: I have my blocking and detection systems report the IP's to an analyser that expands them with Geographical location and registrar data.

Fun facts: North Korea has only one registered CIDR of 1024 IP's. The Vatican has 321 CIDR's with a total of 13056 IP's.