We consider a project in public transport for using prediction and robust optimization in an integrated way. The focus is on integrated vehicle and crew scheduling. A mathematical programming model is provided allowing to incorporate various types of richness. Motivated by the use of electric buses, the number of available vehicles (depending on the battery, driving skills, temperature etc.) is uncertain and may be predicted using appropriate tools. Robustness is included into the optimisation model allowing to consider two types of uncertainty (i.e., uncertain input data being reflected in a given objective as well as uncertainty in the number of available vehicles reflected as parameter on the right-hand side of the model).