Personal Baseball Stats
Name: Kevin McMillin
Number of Teammates Desired: 1
Description of Idea:
Personal Baseball Stats plays off the love many baseball fans have for endless statistics analysis. Baseball fans, scouts, and managers will select their favorite players or prospects at least in part by looking at various statistics which they feel are predictors of future success. Personal Baseball Stats allows a user to input their personal history of attended baseball games and look at how those stats change when constrained by only those events you've witnessed. Who's the best player in the games you've seen? What does your personal all-time all-star team look like? Which teams and players under- or over-perform when you watch them?
Description of Data:
The primary datasource I plan on using is the Retrosheet.org database, available here. It contains a plaintext record of every baseball game dating back to 1948 at the play-by-play level for each game. The data would need a little cleaning, and there's too much to load onto the device so it would need to be culled from a network. Each play is represented using a fairly complex string that can be parsed.
Additional Information: I only picked baseball because it has a particularly rich and accessible dataset, with play-by-play records for every game dating back to 1948. I also know a fair amount about baseball and have attended about 50 games over my lifetime as well, so I would be a pretty good evaluator.
My formal background is in computer science, and a project this data-rich I feel could benefit from a great designer.