sorry to chime in so late here, but i think that the most prevalent areas of this problem are in town. someone or several someones all went around and POI'd groups of every npc in several positions.
the problem isnt in the system implementation, its its use. over the space of 1 day or maybe 2, people found the same guy and wrote down his /loc and all posted. then an extremely overworked POI programmer type just fired them all into the system. and you end up w/ a cluster of 12 of the same three npc's.
another issue i see is the non-necessity to POI every NPC in every city section. some of them give quests yes. but i have no wish to see a POI for Blort the weaponsmith in big bend. there are guards in game that give perfectly good wp for NPC's.
perhaps categorizing and filtering submissions of every "guard #2" ingame is a bit overzealous. there's just too much grey area for people to sift through, this quest but not this.. this funny tree but not this crooked rock. I just cant begin to imagine the paramaters really. but right off the bat id say 3/4 of every in-city poi need to go.
The main question to look at i think is what's the purpose of the map mod.
if its to list and categorize every quest and every npc and every special location in-game. then yes, the total focus needs to be on how to represent them on the map, and make them logically selectable and readable. (to include mulitple overlays for quest, npc, raid, lvl requiremnts, quest start, quest checkpoint, a zoom feature, and/or significant symbology changes.)
but if the point is only to increase the overal intuitiveness of existing maps, and install dungeon maps where there were none... then that's a whole different issue.
Perhaps limiting the mission statement slightly, or generating one at all, will increase the focus and effectiveness for your efforts. It seems to me you've [well we've, i helped plot a map
] got an abundance of motivation and effort and teamwork, w/ no solid goal in mind.