So, I rather like data, what do you think I should collect to have my own little offline library?

javascript frameworks, you'll never run out of things to collect!

Write a spider.
Download everything.

That really depends on what the purpose of your offline library is. If you want general info, Wikipedia has a downloadable offline edition that might be a good start. There are also online repositories of books on certain subjects that you can download. Personally I buy paper copies of any ebooks that I want to have reliable access to when the power goes out, because I live in a rural area and our power does go down from time to time.