With the rise in traffic congestion and associated costs, it becomes crucial to readily make available accurate traffic reports to the general public and also to predict the traffic levels to mitigate further congestion. Various tools and technologies have emerged to solve the aforementioned problem, which comprises of secure and accurate data collection, storage, utilisation of this data for the purpose of prediction, and making required data available to the public. Motivated from the aforementioned discussion, in this paper, various approaches to solve this larger puzzle have been discussed and analysed, and a holistic solution combining the power of blockchain, InterPlanetary File System (IPFS), and neural networks has been suggested. Simulation results show that an LSTM model with 50 time steps and 200 units in the hidden layer, followed by a dense layer leads to minimum Root Mean Square Error (RMSE) value, with a randomly generated but complete dataset. Security analysis of the proposed solutions shows its efficacy compared to state-of-the-art approaches.