The goal of the project was to develop a simple yet rigorous method for mapping Fuel Poverty across a city or region. By using proxy indicators from the U.S. census, we were able to create a Multivariate Gaussian Model capable of predicting Fuel Poverty. The model was trained with data from the 167 block groups in Worcester MA and could, in general, predict Fuel Poverty fairly accurately. However, further research is needed to better evaluate the model226128159s effectiveness, and expand model training and the number of proxy indicators used.
Worcester Polytechnic Institute
Environmental and Sustainability Studies
Major Qualifying Project
All authors have granted to WPI a nonexclusive royalty-free license to distribute copies of the work, subject to other agreements. Copyright is held by the author or authors, with all rights reserved, unless otherwise noted.