Community Projects

The Data and Decision Sciences Lab is constantly working with local community partners on several data related projects that have a local and global impact.

A Ratio-Based Method for Predicting Point Differentials in Sports

Year: 2015
Team Members Involved

The application of statistical models has been used extensively to predict the outcomes of sporting competitions. However, for predicting point differential, the existence of a single model that is easily transferable across multiple sports is lacking. A primary reason exists because each sport can be defined by a unique set of characteristics. A characteristic that they all have in common though is that, for any given team, we can express their score as a fraction of the total score. This project propose a family of models based on that idea which can be easily transferable from sport to sport. For an initial application, the project fitted the data from the 2014 NFL regular season, using the 1970-2013 seasons as historical baselines. This project has also been published as a Master’s thesis.

Mappings of the NearestRounded Method of Fitting Cut/Instance
Potential Improvements on the NearestRounded Method of Fitting Cut/Instance