The MacMillan Dictionary blog has always got good stuff for English Language students and this post is particularly interesting if you're looking at Language and Gender on ENGB1 or Language Variation for ENGA3.
They link to a fascinating project at the University of Trento in Italy which analyses the lexis of Twitter tweets and breaks them down into gendered usage. So, you can search for Robin Lakoff's famous "women's language" features - words like adorable, gorgeous, lovely, divine - and even two word "compounds" - really gorgeous, so cute - to see which gender uses them more in their sample. They've also got a collocate tool that allows you to look for the most common words associated with other search items. So you could look for "pie" and find that on of its male users' most favoured collocates is the Homer Simpson-esque "pizza pie" (mmm... pizza) while one of the favourite female collocates is "sweetie pie".
Of course, all the usual caveats apply to this kind of broadbrush approach. We don't get much context, we don't get much sense of what types of men and women are tweeting and we don't get to see raw data to explore it further, but all in all it's an interesting approach and worth having a look at.