Predicting the energy consumption of a commercial location through machine learning

Abstract

The client has need for acurate predictions of power usage for their assets. There is historic data for all of their assets, they want to use it to predict the power usage. They need an application that can quickly predict the power usage for an asset. The project required knowledge about Machine Learning, which we were unfamiliar with. Another challenge for us was that we both had never worked with Python before. Our research involved getting to know how what kind of machine learning methods there were and how the different algorithms worked. We used an agile development strategy, scrum, to incorporate the wishes of the client. For the implementation side of the project we used Test Driven Development. Our product is an application that will be accessed via the commandline. The results will show metrics in how the algorithms have performed on the dataset. All of our core features and the requirements have been implemented.