About the Project

WillowWatt is partnering with Willow to improve how organizations understand and manage energy usage across large facilities. Buildings in the U.S. consume 76% of all electricity and produce nearly 40% of total CO₂ emissions, so gaining clearer insight into consumption patterns is essential for reducing environmental impact.

Our capstone project focuses on forecasting campus-level energy usage and identifying periods of highest demand so that facility managers can make informed decisions. Using Willow’s digital twin platform and historical energy data, we are developing a system that highlights upcoming peak load periods, giving teams greater visibility into when strain on the grid is most likely.

We’ve built and trained a machine learning model using Random Forest Regression to generate short-term energy forecasts. The model is developed in Python with Scikit-learn and exported in ONNX format so that it can be deployed within Willow’s ecosystem. Our tool visualizes these forecasts, flags predicted peak usage windows, and provides insight into overall demand trends throughout the week.

By presenting clear and actionable forecasts instead of requiring manual analysis, this system helps decision-makers anticipate high-load periods before they occur. The long-term goal is to support better planning, reduce unnecessary energy costs, and contribute to sustainability efforts across campus operations, aligning with NAU’s carbon neutrality goals for 2030.

The initial concept for this project was provided by our sponsor, in the form of a Capstone project proposal, but the resulting system reflects months of collaborative design, prototyping, and refinement with direct input from Willow's Director of Energy Transformation.