While I find the idea interesting, I am doubtful there will be any ubiquitous technique that every or even most networks will be willing to implement for this purpose, especially since not every prefix is likely to be presented from a single geo; some operators likely have one prefix broken into smaller networks, each of which are in disparate geos while being presented to eBGP/DFZ as an aggregate prefix because, you know, they are trying to be good netizens by being mindful of RIB and FIB limitations.

Even if some sort of self-applied geo-tagging of prefixes were defined clearly by some body like IETF or something, it would likely take a tremendous amount of motivation to get operators to actually implement such a thing and then also keep it up to date over time. We can't even get people to do the sensible BCP38 filtering on a large scale and that has pretty significant benefits whereas I, at least at first glance, don't see any noteworthy value of such a geo-tagging mechanism to provoke active implementation by the majority. Maybe something could be embedded in BGP origin validation rpki... like stuff some additional geo bits stuffed into that. This is not a thoroughly thought out idea, just kind of tossing it out there as it is something that is gaining some momentum and sits in the same realm of administratively assessing BGP prefixes.

You might consider something fully within your own administrative control like RTT testing to IP addresses on various prefixes to gauge an approximate distance, though I suppose this idea would be severely unreliable due to many variable conditions affecting RTT that have nothing to do with geographic distance. ¯\_(ツ)_/¯

Best of luck on your search