collecting, by a computing device, information for a future vehicle trip that will be made by a user, the future vehicle trip corresponding to a particular pair of origination and destination points;

generating, by the computing device, multiple routes for the future vehicle trip, the multiple routes all corresponding to the particular pair of origination and destination points;

predicting, by the computing device, multiple amounts of fuel consumption for the multiple routes respectively by providing the multiple routes for the future vehicle trip to an artificial neural network that is trained using one or more sets of training data that are adjusted based on a modeling function, wherein the modeling function comprises at least one of a loss function or a cost function, wherein for each route of the multiple routes, an amount of fuel consumption by the user is predicted based at least in part upon information for the route and one or more user driving features of the user, wherein the one or more user driving features comprise one or more driving maneuvers each comprising a respective type, a respective speed, and a respective duration, and wherein the one or more driving maneuvers are each configured to be classified by a level of fuel consumption based on the respective type, the respective speed, wherein the respective duration of each of the one or more driving maneuvers, and wherein the respective type for each of the one or more driving maneuvers comprises a respective one of braking, acceleration, speeding, or cornering; and

selecting, by the computing device, a selected route from the multiple routes based at least in part upon the multiple amounts of fuel consumption, the selected route corresponding to a least amount among the multiple amounts of fuel consumption.