Resource management is crucial in many fields.
In this project weather and air quality data (NOOA, CDIAC, EPA from 2006--2013) were integrated to predict the frequency of emergency department (ED) visits (from 2006--2013) for asthma (NYC Health data).
The gradient boosted tree model (precision|recall ~83%) performed better on hold-out data (all visits in 2014) than random forest and logistic regression models and is used to render a prototype calendar web application for prospective ED resource management in D3.js.
A summary of this assessment can be found here.