Abstract: The present disclosure relates to a recommendation engine and a method for providing recommendations during navigation of the vehicle on a road. The recommendation engine provides recommendations in real-time to one or more occupants of the vehicle. The recommendation engine receives data related to the vehicle and environment surrounding the vehicle in real-time. The date along with a historical data set regarding the road is used to generate a training set corresponding to a plurality of conditions of the road. Thereafter, the training set is used to create a multi-variate regression model. The multi-variate regression model is applied on the received data for detecting a condition of the road during navigation of the vehicle. The detected condition of the road is used for providing one or more recommendations to the one or more occupants of the vehicle.