Project Overview
Mass Transit Railway (MTR) is one of the major public transport systems in Hong Kong. The saturated usage and incidents sometimes cause turbulence and delays to the metro network. Currently, when facing disruptive circumstances, the MTR network depends on the operation team. Manual evaluating a large sum of data under pressure is difficult, even for a highly experienced team.
Therefore, we applied neural network learning techniques on MTR's raw data to create a forecast model for train journeys. The model is to provide objective and accurate analysis to train operators on the real-time railway situation. Hence, when a delay occurs, they can act early and accurately to reduce the amount of delay time suffered.
In this project, I am responsible for preprocessing the data collected from MTR, developing the train movement replayer, and training the baseline models (ie. Linear Regression, Decision Tree, Random Forest) for result comparison.