###SportsCanary uses data from Twitter to generate "crowd-sourced" predictions for major sporting events. Currently supports MLB, NHL and NBA. #####As of 04/22 we have correctly predicted 64.865% of NBA games, 46.73% of MLB games and 55% of NHL games.
##Overview There are four parts to this repo:
- Website contains our MEN stack (Mongo, Express, Node) website.
- Twitter_Utils contains our python backend responsible for gathering data for the days games and predicting who will win.
- Gnip_Client contains scripts for interacting with the GNIP search API as well as interfacing with the IBM Watson Tone Analyzer.
- Gambling_Utils contains our proxy as well as code for interfacing with the BetfairAPI (which expects a non US connection due to US gambling laws.)
- Eternal_Utils mostly contains scripts for retroactivley updating our database, instrumenting our code, and sending encrypted data.
##Tools Used
Data: Twitter, GNIP Search
Database: MongoDB
Website/REST Endpoints/Search: NodeJS, Jade, Elastic
Data Gathering Logic: Python