Forecasting Binghamton University’s Future Electricity Consumption Using Building Classifications and Historical Weather Data as Inputs into a Machine Learning Model

With the rise of energy consumption intensifying the adverse impacts of carbon emissions, different alternatives to achieve sustainability and carbon neutrality are explored in academic conversations. Energy consumption forecasting, which refers to analyzing historical electricity usage to predict future trajectories, is a promising solution to achieve these visions of sustainability as this methodology enables organizations to identify their energy wastage to decrease their electricity usage, while also having the additional benefits of creating energy budgets and saving costs (Amber et al., 2017). My research specifically focuses on creating an energy consumption forecasting model for Binghamton University, using data from 2017-2020. To engineer this model, I will be classifying different university buildings and analyzing the correlation of these building’s electricity usage patterns with historical weather data. This information will then be processed through a machine learning model using AzureML Studio to forecast future electricity usage patterns. By presenting a model for forecasting future energy consumption utilizing building classifications and historical weather data as inputs, my research will illustrate the accuracy of energy consumption forecasting. I specifically argue that Binghamton University should implement this energy usage forecasting model as they have the potential to reduce carbon emissions, help financial budgeting, and minimize energy expenses.https://orb.binghamton.edu/research_days_posters_2022/1000/thumbnail.jp