Date of Completion

2015

Document Type

Honors College Thesis

Department

Statistics

Thesis Type

Honors College

First Advisor

Richard Single, Ph.D.

Keywords

energy, prediction, data, R

Abstract

This paper examines energy usage patterns of residential and commercial households in the state of Vermont. There are three main objectives of this paper. The first is to aggregate the data from the “high bill” calls that the Vermont Energy Investment Corporation (VEIC) receives. This is accomplished by cleaning and organizing the data into one Excel spreadsheet using Google Refine. The second objective is to predict the presence of an appliance based solely on the total usage of a home. This is accomplished using data collected from 24 high performance homes in Vermont. This data set is analyzed using histograms, correlations, clustering, random forests, tree plots, and heat maps. The final objective is to predict the future electric usage for the commercial accounts in Vermont. This is accomplished by analyzing six different models on the monthly usage from the commercial accounts and then comparing the results of the models to determine the best approach.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Download

Share

COinS