Of course, tying the players in is important. To an extent, that depends on the gaming group. I'm fortunate that in my group I have one player (usually playing fighters) who is happy to go in swinging his mace, picking low INT characters and not caring for much else, while I have another player who just starts tying his character in on his own and takes over the lion's share of the world-building that makes the characters click (he literally invents taverns, quests, locations, foes, possible twists, hooks - and I just steal those ideas I like and run with them). That way, while I retain authorial control of the world, the characters are in a sense pre-inserted ... example: they come into a town ravaged by a dragon, and his necromancer nods wisely and says, "indeed, that is the dragon we enraged when we tried to steal his blood to make a love philter for that siren that will otherwise not wed me. It is obviously coming for us and we must stop it." The daemon summoner is of course enthusiastic about using the dragon's wing membranes to make a light armor for himself, the ice wizard comes along because, well, it's his town, and the fighter is just happy to be fighting a dragon . I'm not sure about the dwarf's motivation, I think he wanted to eat a dragon steak, so he could boast to other dwarfs. Dwarfs are strange in my setting ... almost as strange as elves (nasty changeling shapeshifter basterds).

I just think the best way to get the players invested is to discuss things with them and be pretty open about what they can get out of a region (love, money, castles, mills, titles, whatever). As a DM I personally don't care how much non-tangible (i.e. not related to in-game mechanics) power a player character accumulates, since they can't take that into a dungeon anyway. Also, I run my games pretty hard-core, so I've often had players happily retire their characters from adventuring once they bought a nice inn, married a princess or other things ... better to become a living legend in your little fantasy town than another tombstone in the Valley of Heroes.

Sorry for the excursion ...

1. Locations - I've always found that locations can easily be recycled. For a sandbox area, you don't need all the locations. You need some locations - and then place them as needed. Other locations (if you run out or something) can be declared destroyed, snowed in, inhabited by something too dangerous, whatever. Basically, locations are modules which, unless they're relevant to the main story, you slot together with modular opponents and modular treasures. That also lets you actively recycle "on the fly" ... I'll try to explain better as I prepare the module.

2. Multiple paths - absolutely. And multiple outcomes. I plan to create a time-line of events that happen. Now, some of the events (e.g. weather, earthquakes, invading forces, whatever) are external and out of the players' control. Others are player-decision dependant. If they talked the lord out of stealing all the peasants' food, the peasants don't starve, bandits don't start appearing, etc. If the lord went and stole from another lord, you get a war starting. If the players do nothing and all hell breaks out, with the peasants barricading themselves in, you start getting undead ice-wights and what not roaming around the land due to the surfeit of unburied and unsanctified corpses. Etc. etc. I plan to create just hints of the main story at the beginning, then offer different potential outcomes for the players - from defeating the villain, just escaping the valley and leaving it to its fate to joining forces with the arch-villain, possibly becoming ice-wights themselves, etc. - which will just mean there's a lot of potential options for the next cycle of adventures!

I'm often a very anti-rail-roady DM

*though instant death traps are fun