@jpreisendoerfer @factotum I know it's not trivial, which is part of why I use the more or less neighborhood precision of the #LocationLabels to do a rough match on a meaningful level, as in you can match things to be close enough and weed out the differences in a small post correction. You don't need it to be perfect, you just need it to be computationally feasible with the predicted amount of data. If you can contain location data in small enough boxes to figure it out one by one, that is good enough in practice. Which is also why I think that you cannot be format-agnostic here, you need to make a choice for the underlying format that is going to be the norm, and most formats CANNOT handle this. the #LocationLabels idea was come up with exactly to have a possible, even if rough, solution for this problem that is good enough for practical social network usage.