A computing system for predictive traffic management using virtual lanes. In an embodiment, the system dynamically monitors and collects traffic conditions in real time, performs analytics on the collected traffic data, utilizes a neural network or other self-learning computer to assist in predictive traffic modeling, and interfaces with a public transfer system to provide an allocation/reallocation of lanes available for traffic use to optimize traffic flow and/or control traffic signals, and can provide vehicles (human driver or driverless/self-driving) with real time optimal route guidance, including use of alternate routes and a holographic image that shows and may also provide audio indications of lane allocation.