Electricity Production Prediction by Microsoft Azure Machine Learning Service and Python User Blocks

Abstract

In this chapter, the forecasting of electricity consumption and production is conducted by analyzing indicators from previous years. The problem is addressed using machine learning within Microsoft Azure Machine Learning Studio. The outcome is an independent service integrated into Excel, enabling consumption forecasting for specified dates. The Excel user interface is developed using Visual Basic for Applications. Python was used to create user blocks for modifying input data pools and forming graphical dependencies, seamlessly integrated into the original modules of Microsoft Azure Machine Learning Studio. An additional aspect of the forecast results involves evaluating the quality of the predicted electricity consumption indicators. The materials used for this chapter were sourced with the support of Ukraine's National Power Company UKRENERGO.