I'm currently involved in a project to design and implement an in-house question-and-answer community for our university. We've explored some open source frameworks such as OSQA, but we are also considering designing the community from scratch.
Thus, what is the database structure of Stack Overflow? Or, how should the core database structure of our system be?
You can start by downloading the public data dumps and / or playing around with the Stack Exchange Data Explorer. This answer has all the details you'll need on the schema for both. There are some differences from the actual schema, for example votes are anonymized so there would be no point to share schema details that tie votes to voters. But I'm guessing it's already a lot more than you would have imagined it'd be available. Checking out the API is also worth your time.
Other than that there are always bits and pieces of related information on the blog and on various questions here on Meta, when the developers feel like sharing such information on feature-request or bug questions.
If you build an amazing Q&A platform and you happened to benefit from all that freely available information and from Stack Overflow in general, please consider releasing the project as open source so others can benefit from it. Good luck!
