Sources, APIs and techniques for academic data-mining

So I'm gonna do a historical trends analysis for one of my education literature review class. I'm considering use the abstracts and keywords from the peer-reviewed journals since 1970s, I know Springer has a开发者_如何转开发n API which might be useful, what other sources/APIs are there? I never did any data-mining and analysis before, I know a bit about ruby and less about python. Any other advices are also welcomed.

I think you can get abstracts of medical publications (not the fulltext though) using the PubMed API. That'll give you something on the order of a million documents.