I'm thinking about setting up some machine learning algorithms, particularly to use with nba and college basketball prediction. I did a couple of them last year for a class with the men's NCAA tournament and it worked quite well, so I'm interested in keeping it up. Problem is, data collection was a pain because I did it all by hand, copying and pasting. Setting up a script to do it seems annoyingly complex as nba.com and espn.com boxscores have tons of crap in every webpage. Does anyone have any better ideas? Know any sites that do this data collection already? Thanks.