A Twitter bot that watches MLB games (Indians games, in my case) by polling the Gameday API and tweeting alerts during high-leverage game states.
Based on, but not associated with, Leverage Index calculcations by Tom Tango found here: http://www.insidethebook.com/li.shtml
For myself and many sports fans in 2016, a big part of our experience, be it live or at home, is the second screen. And in particular: Twitter. The live-feed nature of Twitter is a great complement to live sports. I myself maintain a Twitter list of people who tweet interesting commentary during Indians games: the team account, all the beat writers, some bloggers, and even a few interesting Twitter bots. Following that commentary during an Indians game makes my fan experience richer.
In sabermetrics, there is concept called leverage. During a game, some situations are more critical to the outcome of the game than others. A tie game with a runner in scoring position in the ninth inning is more critical than a tie game with a runner in scoring position in the second. This pressure is quantified in a statistic called Leverage Index. As fans we can intuit leverage roughly, but some situations are deceptive. Wouldn't it be great to know in real time if a situation is high leverage?
So we made and deployed a Twitter bot, @cleLeverageBot. It polls MLB's Gameday API during Indians games, and if a particular situation meets a leverage index threshold, it posts a tweet: Heads up! This situation is high-leverage, and what happens right now might strongly influence the outcome of game. You probably want to be paying attention!
Future development: the aggregate Leverage Index data will be fed to a website, baseballtribe.com, where fans can see how exciting a game was, how exciting future games might be, etc.