Games built around a school setting have a bunch of advantages. There is a strong scaffold for goals and rewards, there are clear social relationships and subcultures, the choices are easy to understand but still important, there is a good place to hang your statistical gameplay, people all start at a similar power level, and it's appealing to people who have been to a school and like to see echoes of that, except better.
The obvious example of a school setting is Hogwarts, and the "magic school" setting is so widespread it's gone from being a subgenre to almost being a genre in its own right. Hogwarts certainly wasn't the first, but it was probably the popularizer. Dozens of comics, anime, and games are based in schools and benefit from many of the advantages listed above.
The big innovation Hogwarts popularized was the idea of houses. That feeling of being lost or adrift that you may have felt during school is gone. You instantly belong, you instantly know the kind of person you're expected to be and the kind of things people will look for in your actions. If you value pushy heroics, join Gryffindor and they'll value your personality. If you like considering and understanding things, join Ravenclaw, that's where they'll value your personality. Just create a house for each kind of audience member and your audience will project themselves into your universe faster than you can blink!
In games this is also quite an opportunity, allowing players to instantly tell you what variant of your game they want to play. But, oddly, most games don't use factions that way. They use factions to give statistical bonuses, which is totally dropping the ball. Factions such as school 'houses' should be there to give the players a framework of valuation, not a particular bonus and penalty. You can be a smart Gryffindor or a mighty Ravenclaw, both are perfectly acceptable. It's just that your actions would be rewarded based on your houses' preferences. You can be a smart hero, a muscular buddha.
This is the strength that games leave at the door. The opportunity they throw away.
Also, if you make it into an open game where players can live their student life as they see fit, you need to have a more heavyweight backbone of adventure and progression than most school-based games use. Of course, that's because actual school doesn't exist to let you adventure or choose your path. Real schools exist to teach you a million things you need to learn to work at a factory, and then kick you out. So you'll have to build a framework that goes much further than simply 'classes and grades'.
Let's posit a game, just for the sake of being clear.
This is a school for scientists and makers, rather than magicians. The reason? Well, fundamentally science is about discovering stuff and making is about creating stuff. That'll form the backbone of our adventure and progression system, with classes forming the backbone of our basic goal and reward scaffold.
For houses, rather than use different personality types, we'll use different kinds of cooperation. We'll play a bit fast and loose with this. Rather than actually talking about different kinds of cooperation that scientists use, we'll instead codify it as famous science projects that have different kinds of behavior within them.
This was based on this G+ post via this G+ post. The examples are all from astrophysics, but I would argue that this isn't bad. Astrophysics is full of projects which are big, concrete, and open. You can see how the pieces work, they are actual pieces that keep working for years or decades, and they work in specific ways to get science done. These sorts of things could form the basis of our house system, reflecting how you would prefer to acquire knowledge and build the future.
For example, the LHC is about slamming particles together as hard as possible, generating so much data that it takes weeks to filter through it. House Hadron students might therefore be expected to perform big experiments and then spend quite a bit of time poring over every detail of the experiment. The makers would, for example, build a complex 3D printer and then run it into the ground by stress-testing it for ten weeks.
On the other hand, House Curiosity is about observations in places people can't reach or rarely go. Curious students would be expected to search caves, mountains, the bottom of the pond out front, the steam tunnels, the headmaster's locked desk drawers... anywhere people don't ordinarily go, Curious students would go - looking for inspiration and information. A Curious maker might build a submersible to explore a pond, or a quad-copter to fly to the top of a mountain.
Once you get over the extremely geeky nature of the game, I think you can see that there's meat here. Players would be able to decide pretty quickly whether they would prefer large projects or exploration. Or any of the other houses that get put into the game.
Moreover, the fact that the students are gathering data, investigating, and building means that they have a scaffold for adventures and growth outside of the classroom, whereas in a magic-based game you might have to script all those adventures line by painful line.