There are a ton of people who are more knowledgable about college basketball and statistics than I am, who have tried, and weren’t that successful at predicting the outcome of the NCAA tournament.
I’m going to try anyway. At the very least, I’ll get practice with Node, Firebase and writing clean, readable code. If, by some chance I get a formula that works, I might be able to get $1 billion dollars out of it
This was sparked, primarily, by me being sick of how unpredictable my Tar Heels have been this year. Surely there is something that’s “causing” them to lose to borderline-division-2 schools, and beat three top-ten schools all in the span of a month?
so, DATA. So far in this project I’m using:
Firebase – cloud-based db and app hosting. I might switch to a neo4j graph database later if it turns out there are too many joins to want to deal with)
CasperJS – ‘dom-liberator’ (but also used for testing)
Step one: Set up Node server and Firebase DB.
Firebase makes it really easy to get your DB set up in a Node server. Literally two lines:
Step two: set up getting data from a certain sports property’s API.
This sports property makes some, not very useful, information available on their free API. What they do offer for free is a URL to the player’s pages, which can be easily parsed to get their game logs. This was a simple GET request, posting the URLs with player names directly to my DB.
That’s where I am so far. Next, I’ll be using Casper to get data from the game logs I saved earlier.