The eVIP (Energy Visualisation Integration and Prediction) project aims to predict the load curve of electric vehicles in a semi-private context related to hotels and restaurants. Using the gradient boosted tree algorithm, it is possible to predict the consumption of a hotel with an accuracy of approximately 83.8% with nonintrusive devices. By using this prediction and the data collected when an electric vehicle is being charged at the hotel's charging station, the peak consumption of the hotel can be optimized.We have also opened the way for V4G (Vehicle for Grid) to allow bi-directional energy flows in this semi-private micro-grid and propose flexibility services to Distribution System Operators (DSO).