This paper highlights the extension of PI analytics functionalities taking advantage of the power of Machine Learning and Big Data Analytics. These techniques have been applied to the forecast and reduction of the energy consumption in an upstream plant. PI system features such as high availability, smart data interpolation and cleansing capabilities are leveraged along the overall pipeline of any machine learning product development: exploration, training and deployment. The final model is deployed in the Big Data Infrastructure and fed with near real time data. This solution achieves two main results: a reduction of co2 emissions and a smooth transition for experienced plant operators towards advanced Analytics solutions and data science techniques in the Oil & Gas Business.