Video by Hilary Mason: An Introduction to Machine Learning with Web Data
oreilly.com
oreilly.com
Also good advice from the video: ask politely for data. About 12 years ago, I wanted access to a large corpus of news stories and none of the national news web sites I contacted would respond to my requests for legal access to their data (i.e., scraping their sites). I then contacted a local paper we subscribed to (Arizona Republic), told them I was a customer, exactly what I was going to use the data for, etc. I received a permission letter fairly quickly.
And let me know if you have any questions!