Industrial loads in non-residential buildings have significantly contributed in total energy use throughout the world. This paper aims to develop a data-driven risk-based framework to predict and optimally control industrial loads in non-residential buildings. In the proposed framework, first, a set of predictive analytics tools are employed to identify the patterns of industrial loads over time. This also includes a high-dimensional clustering model to allocate industrial load profiles into smaller groups with less variability and same patterns. Once the patterns of industrial loads are identified, then a classification model is implemented to estimate the best class that matches with any new load profiles. Ultimately, the proposed framework provides a risk-based model to calculate and evaluate the total risk of energy decisions for the next day. This is coupled with a utility function structure to help decision makers to take best demand-side actions. The efficiency of the proposed model is investigated through a real world use case. (C) 2015 Elsevier Ltd. All rights reserved.